In recent years, with an increase in advanced-age pregnancies and the application of assisted reproductive technology, the incidence of twin pregnancy has been increasing worldwide and in China [1, 2, 3]. However, twin pregnancies are high-risk pregnancies that are prone to be complicated with maternal and foetal complications. Currently, there are few twin studies with respect to Chinese tertiary hospitals. This study was designed to perform a descriptive analysis of twin pregnancies in a Chinese tertiary hospital, summarize the clinical characteristics of twin pregnancies, lay a foundation and indicate key directions for clinical diagnosis, treatment and further scientific studies.

In recent years, the incidence of twin pregnancy has shown an upward trend worldwide and in China. The recent twin birth rate in the Chinese population was 2.61%. From the perspective of epidemiology, twin pregnancy has become more common in the current obstetrical clinical management.

In this study, monochorionic twins accounted for 28.73%, and dichorionic twins accounted for 71.27%. According to previous data, the incidence of monochorionic twins in natural pregnancies was approximately 1/250 [10], accounting for 20% of all twins. However, with the application of assisted reproductive technology, the current incidence has increased 2–12 times [11]. The rate of monochorionic twins in the present study was different from that in previous reports, and different nations, races and sample sizes may account for this discrepancy.

In addition, in this study, 717 women were 29.35 $\pm$ 4.78 years old, women between the ages of 25–34 years accounted for approximately 70%, and approximately 70% of the women were experiencing their first deliveries. These results were somewhat different from previous results. Otta et al. [12] believed that the incidence of twin pregnancy increased with the age of the mother, with dichorionic twins being particularly common. Kullima et al. [13] found that 60% of twin pregnancies occurred in the maternal age group of 20–29 years. Aduloju et al. [14] showed that more than 50% of mothers had delivered multiple times previously. Some of the reasons that may account for these differences include diverse birth conditions in different countries and regions, varied sample sizes among different studies, and exclusions of some factors, such as the application of assisted reproductive technology.

Twin pregnancy is a high-risk pregnancy, which increases the risks of obstetrical complications. In this study, preterm birth, hypertensive disorder and PROM were the three most common complications, with incidence rates of 60.39%, 28.45% and 21.90%, respectively.

In the 717 twins, the gestational weeks at delivery were concentrated between 34–38 weeks (72.67%). According to the 2013 birth information published by the National Center for Health Statistics (NCHS), the delivery weeks of singleton births were focused at 37–40 weeks (75.91%) [1]; thus, it appeared that there was a large difference in the delivery week between twins and singletons, which seemed reasonable. Twin pregnancies were more likely to deliver earlier.

However, the rate of total preterm birth in the study population exceeded 60%, of which 75% of the cases were iatrogenic preterm births, 10% occurred before 32 weeks, less than 20% of the cases occurred at 32–33${}^{+6}$ weeks, and 70% were at 34–36${}^{+6}$ weeks. In 2013, 132,324 twin pregnancies were delivered in the United States, of which 56.56% were preterm births, and 11.31% of the cases occurred before 32 weeks. This group of data was similar to the results of this study. These two sets of data corresponding to singletons in the USA in 2013 were 9.71% and 1.54%, respectively [1]. Compared with singletons, twin pregnancies had earlier delivery weeks and a higher incidence of preterm birth; preterm birth is definitely an important reason that increases the risk of neonatal morbidity and mortality. The common investigations to prevent preterm birth in twins included cervical cerclage with ultrasound monitoring, progesterone use, cervical pessary use, bed rest and uterine monitoring. Currently, there is a lack of effective, evidence-based interventions for the prevention of preterm birth in twin pregnancies [15].

Hypertensive disorder is the main threat to the health of mothers and foetuses. Hypertensive disorder was demonstrated to be one of the three most common complications, with an incidence rate of 28.45% in the current study. However, the incidence was higher than the incidence rate of 22% in a previous literature report [16]. Some reasons may account for the discrepancy. First, there were four types of hypertensive disorders included in the current study, while only pre-eclampsia and gestational hypertension were included in the previous study. Second, our hospital was a tertiary organization that managed pregnant women with hypertension, especially those with severe conditions.

Pre-eclampsia accounted for the largest proportion (nearly 86%) of hypertensive disorder types. Among pre-eclampsia, most cases were of late onset and severe type. According to a previous study by our group, there were poor perinatal outcomes (such as preterm birth, iatrogenic preterm birth, intrauterine growth restriction, neonatal neurological complications, and reduced birth weights) for twins complicated with pre-eclampsia compared with normotensive twins [17].

According to Fig. 1, we thought that if the pregnancies were terminated before 33 weeks, vaginal delivery was more common. Most of these cases were terminated vaginally or through artificial abortion due to PROM, inevitable abortion, sIUFD, etc. Some cases had emergency caesarean sections before 33 weeks due to pre-eclampsia, scarred uterus, etc. After 33 weeks, caesarean section rates were higher due to obstetrical complications and large baby weight.

Emergency caesarean section rates were higher (approximately 70%) between 33 and 37 weeks, mostly due to the higher incidence rates of obstetrical complications (pre-eclampsia, foetal distress, PROM, threatened preterm labour, etc.) and twin complications (TTTS, IUFD, sIUGR, etc.) in this time period. When the twins survived through that duration, the emergency caesarean section rate decreased, and more than half of cases were terminated with elective caesarean sections between 38 and 39 weeks.

The caesarean section rate has been maintained at a level of 70%–80% in recent years. However, in 2013, there was a ‘low valley’, with a sudden decrease in the rate to 62%. This decrease may be related to an increase in vaginal deliveries (35%) due to PROM and threatened premature birth in that year. The delivery modes of 717 twins were caesarean section, vaginal delivery and artificial abortion, which accounted for 75.87%, 21.06% and 3.07%, respectively. In the study population, caesarean section was the main delivery mode for the majority of women in our hospital, which was perhaps related to the fact that some women were with severe obstetrical complications in this tertiary hospital, or had conceived by assisted reproductive technology to achieve such a desired pregnancy. Choosing an optimal delivery mode for twins is an important issue in clinical management.

In 2013, a multi-centre randomized trial by the Twin Birth Study Collaborative Group randomly assigned 1398 and 1406 pregnant women to a planned caesarean section group and planned vaginal delivery group, respectively. The final caesarean section rates were 90.7% and 43.8%, respectively. Neonatal adverse outcomes in both groups were followed up, and it was found that a planned caesarean section did not significantly change (reduce or increase) the risks of perinatal death and serious neonatal complications for the twins who were born between 32–38${}^{+6}$ weeks when the first foetus was head position [18]. Based on this background, in 2016, the Twin Birth Study Collaborative Group conducted a subsequent study investigating long-term mortality and neurodevelopment in 4,603 children and found no significant difference between the planned caesarean section group and the planned vaginal delivery group [19]. Currently, there is no authoritative evidence supporting planned caesarean section for improving the short-term and long-term outcomes of twin pregnancies. Mostly according to the policies in the world and in China, our choice of delivery mode was comprehensively decided by chorionicity, maternal history, complications, foetal position, local medical care level and clinical experiences and techniques. Vaginal delivery may be attempted more with increasing clinical experience.

For neonatal adverse outcomes, the incidences of 1 min Apgar score $\le$7, SGA, and NRDS were all between 10%–20%, and the incidences of VLBW were 7.06%, which were similar to the data from NCHS [1]. These outcomes mostly occurred before 37 weeks, and 1 min Apgar $\le$7, NRDS, and VLBW were more common before 34 weeks.

Based on Fig. 3, the common features were similar to those in previous research [20], of which monochorionic twins were consistently smaller than dichorionic twins after 27 weeks of gestation. The sample size of the birth weight curves was not large enough in the current study, which could be the main reason for the imperfect curve and the decreasing trend after 38 weeks in monochorionic twins. Of course, there would be some differences in curves among various races and regions.

The incidence of weight discordance exhibited a gradual decrease in the range of 5% to 35%, which was consistent with what was reported in other literature [21]. Growth discordance was demonstrated to be a risk factor for intrauterine death, especially when at least one foetus was small for gestational age (SGA) [22]; growth discordance also increased the risk of neonatal asphyxia [23] and affected the long-term neurodevelopment of the smaller twin [24]. From 20 weeks, discordance in estimated foetal weight should be calculated and documented at each scan. If discordance reaches 25% or more, a referral should be made to a tertiary-level foetal medicine unit for assessment, increased foetal surveillance, including foetal Doppler, and planning of delivery when appropriate [4].

The limitations of the study included the absence of some clinical information due to the retrospective study design and the limited record system (such as the method of fertilization, twin family history, and BMI) and the inability to compare twins and singletons in the same population. Research on malformations of twins from natural conception and in vitro procedures would be welcome.

It should be pointed out that, in the present article, comparisons between monochorionic and dichorionic twins were not presented because these contents are the main text of another published article by our team, which was why we did not show them here. The aim of this study was to show an overview of the population twin outcome, not to provide a comparison. Studies on twin perinatal outcome comparisons between monochorionic and dichorionic twins are certainly meaningful and welcome.

In conclusion, an overview of the baseline characteristics, maternal outcomes and perinatal outcomes in the Chinese twin population was provided, including 717 cases in nearly ten years, which may be helpful for obstetrical management. Obstetricians should pay greater attention to the clinical management of twin pregnancies to obtain better maternal and foetal outcomes. This study provided more information for twin clinical studies from the perspective of the Chinese twin population.

YT reviewed the relevant literatures, organized and wrote the manuscript. LXL and HZ conceived of the study, designed of it and provided some part of data. ZT and WYY accomplished parts of figures and tables and contributed in writing the manuscript. All authors read and approved the final manuscript.

The research has been approved by the Ethics Committee of the First Affiliated Hospital of Xi’an Jiaotong University (Grant No. XJTU1AF2015LSL-073) and the clinical research project (NO. XJTU1AF-CRS-2015-003) has been set up. It is in accordance with the Helsinki Declaration of 1975, as revised in 2000. Registration has also been completed on the website clinicaltrials.gov (ID: NCT02732717). The written informed consent was unnecessary due to the retrospective, observational study.

Thanks to all of the members in our team.

This study was supported by the Health and Family Planning Commission Project of Shaanxi Province (NO. 2016D063) from HZ and Science and Technology Planning Project of Shaanxi Province (NO. 2020JQ-523) from YT.

The authors declare no conflict of interest.