A retrospective cohort study was conducted at a tertiary teaching hospital. Data on twin pregnancies after vaginal delivery from January 2016 to July 2020 were collected. The protocol was approved by the Ethics Committee of ** Maternal and Child Health Hospital (Y2020013). The eligibility of potential participants was assessed by medical records. Since all information was anonymized and deidentified before analysis, the informed consent was waived. This project has been registered in Chinese Clinical Trial Registry (ChiCTR2000041228).
We carried out the study at the Maternal and Child Health Hospital, a comprehensive university teaching hospital with 200 beds in the maternity ward. The number of newborns delivered was 10,000–12,000 annually, accounting for over 50% of the total births in ** City. The vaginal delivery rate was around 55% for singleton and 15.59% for twin birth. We collected data on twin pregnancy after vaginal delivery and the 1st fetus with vertex presentation. We excluded women and fetuses with 1) Intrauterine death of either one of the twins before the onset of labor; 2) Fetal malformations; 3) Caesarean section; 4) Nonvertex-vertex presentation or nonvertex-nonvertex presentation; 5) Ultrasound diagnosis of twin-twin transfusion syndrome; or 6) Iatrogenic preterm delivery (Medically induced preterm that was commonly caused by preeclampsia, placenta previa, or fetal growth restriction .).
Fetal presentation was determined by ultrasound before the start of active labor  and confirmed at the end of the 2nd twin birth. Fetal presentation was categorized into the following two groups: vertex-vertex (VV) presentation group and vertex-breech (VB) presentation group. VV and VB are the most common presentation for vaginal delivery recommended by the guidelines. On this basis, the present study was primarily aimed to compare the outcomes between VV group and VB group.
Neonatal and Maternal Outcomes
The Apgar score provides guidance for evaluating the condition and prognosis of the neonates. A score of 7 or higher indicates a good or excellent neonatal condition. Moreover, a low Apgar score (0–3) increases the risk of neonatal death. The 1-min and 5-min Apgar score was evaluated independently by two midwives.
For all infants deemed viable, umbilical artery cord blood was routinely sampled and placed on ice by nurses immediately after delivery. Subsequent analysis was usually completed within 10 minutes after delivery . GEM-4000 Blood Gas Analyzer was used throughout the study.
Whether to transfer to NICU was independently judged by two pediatricians, at least one with a senior professional title. The indications to the NICU were: 1) The newborns with longer duration of intrauterine distress or severe asphyxia need guardianship; 2) Premature infants, very low birth weight infants or small for gestational age infants need close monitoring; 3) Hypoxic-ischemic encephalopathy, intracranial hemorrhage, and central nervous system infection; 4) Newborns with acute or chronic respiratory failure, frequent apnea, need oxygen therapy, endotracheal intubation, and mechanical ventilation, as well as respiratory management for various reasons.
Perineal tears were classified into four grades according to the Green-top Guideline . First-grade tears were defined as the injury to perineal skin and/or vaginal mucosa. Second-grade tears were defined as the injury to the perineum involving perineal muscles but not the anal sphincter. Third-grade and fourth-grade tears were defined as the injury to the perineum involving the anal sphincter complex.
Postpartum hemorrhage (PPH) was defined as excessive bleeding (≥ 500 ml in vaginal delivery) occurring within the first 24 hours after delivery . Volumetric blood loss measurement techniques  was conducted by placing a curved disk under the buttocks of the parturient to evaluate the volume of blood loss. Gravimetric blood loss measurement techniques were conducted by weighing all swabs and towels with blood contamination within the first 24th hours . After subtracting the original weight of all swabs and towels, the difference obtained was divided by 1.05 to be the final bleeding volume.
Demographic and clinical information
Information retrieval, including maternal age, pre-pregnancy BMI, parity, gestational age, birth weight, fetal gender from the registers, medical records, and obstetric database was performed. The complications included hydramnion, hypamnion, heart disease, hyperthyroidism, hypothyroidism, intrahepatic cholestasis, placenta praevia, premature rupture of membranes, fetal distress, abruptio placenta, and umbilical cord prolapse. The measurement of birth weight was accurate to 2 grams. Gestational age was based on the first day of the last menstrual period . If a woman was not sure about her last menstrual period, the larger of the two CRLs (Crown-rump length) would be used to estimate the gestational age in spontaneously conceived pregnancies. In vitro fertilization-embryo transfer (IVF-ET), gestational age was determined by the date of embryo transfer . Preterm infants can be classified into early preterm birth (28 ~ 33+ 6 weeks), late preterm birth (34 ~ 36+ 6 weeks), and full-term birth .
Management of Vaginal Delivery
The Selection of the Planned Delivery
For uncomplicated, dichorionic diamniotic twins, with an estimated fetal weight of exceeding 1,500 g, and the leading twin with vertex presentation, it was suggested to have vaginal delivery. Nonvertex-vertex presentation and nonvertex-nonvertex presentation were scheduled for elective cesarean deliveries. After discussing the advantages and disadvantages of vaginal delivery, patients decided whether they were willing to try vaginal delivery. In case a woman was admitted to the hospital at the end of the first stage or at the second stage, an emergency vaginal birth was conducted.
Management of the First Stage of Labor
A full range of personnel and equipment were available. The personnel for a twin delivery included two obstetricians, three obstetrician-midwifes, an anesthesiologist with an assistant (if necessary), a neonatologist, a neonatal nurse and a qualified sonographic assistance. Both twins underwent continuous cardiotocography during the entire process of delivery using a Polar heart monitor with dual-channel capability. A portable ultrasound machine was available to check fetal presentations. The delivery room was equipped with complete emergency cesarean section.
Management of the Second Stage of Labor
The umbilical cord was clamped and cut after the delivery of the first fetus. After delivery of the 1st twin, the monitoring of the 2nd twin was continued through physical examination, ultrasound, and electronic fetal monitoring. Ultrasound was adopted to assess the 2nd twin's presentation.
Amniotomy of the 2nd twin was not performed until after the head was engaged. The specific finite interval was set to be completed within 30 minutes between delivery of the 1st and 2nd twin. If the 2nd twin was in a breech presentation, we preferred breech extraction if there were no contraindications to this procedure [21, 22]. Pinard maneuver assisted the delivery of the buttocks, trunk, and lower limbs. Loveset maneuver facilitated the delivery of the arms and Mauriceau-Smellie-Veit maneuver assisted the delivery of the head.
After delivery, immediate care included drying the newborn, removing the airway of secretions (if necessary), and providing warmth. For newborns with greater probability to need resuscitation, a skilled neonatal resuscitation clinician took care of the high-risk infant in the delivery room. In our hospital, all health care providers took care of newborn infants (neonatologists, midwives and anesthesiologists) received training. The resuscitation procedures followed Chinese neonatal resuscitation guidelines (CNR, 2016) based on American Academy of Pediatrics Guidelines for Neonatal Resuscitation 2015. The procedures included initial steps in stabilization, ventilate and oxygenate, initiate chest compressions and administer epinephrine and/or volume applied primarily to newborn infants transitioning from intrauterine to extrauterine life.
The sample size was calculated by the primary outcome of Apgar score where 1:2 matching was used for VV and VB groups according to the proportion of the two presentations [3, 4]. Two-sided level of significance of 0.05 and power of 0.8 were set. The final planned total sample size in VB group was 70 and 140 in VV group. The sample size calculation was performed on PASS software, version 14.0 (Number Cruncher Statistical Software).
SPSS for Windows, Version 25.0 was used for data analysis. Continuous variables with normal distribution were expressed with mean and standard deviations and analyzed with two independent samples t-test. Chi-square test was used for categorical variables. Apgar scores over time were compared by repeatedly measuring ANOVA.
Considering the imbalanced baseline and confounding factors, we performed regression models on the maternal and neonatal outcomes, including the multivariate linear regression and binary logistic regression analysis. P < 0.05 was considered statistically significant.