This multicenter study aims at assessing the natural history of twin pregnancy and developing a machine learning-based algorithm to predict clinical outcomes of twin pregnancy during pregnancy and delivery and to determine management strategies that are associated with best maternal and neonatal outcomes. This study will include at least 12 centers from different countries that present at least Europe, South America, Asia, and Africa. Data will be retrospectively collected from January 1st, 2010 to December 31st, 2019.

Twin pregnancies carry higher risks of maternal, fetal and neonatal adverse outcomes compared to singleton pregnancy.They are associated with increased perinatal morbidity and mortality, anemia, pregnancy- induced hypertension, increased incidence of cesarean section (CS), postpartum hemorrhage, prematurity and low birth weight and Increased rate of perinatal death.

This multicenter study aims at assessing the natural history of twin pregnancy, and developing a machine learning-based algorithm to predict clinical outcomes of twin pregnancy during pregnancy and delivery and to determine management strategies that are associated with best maternal and neonatal outcomes.

Medical records of eligible women will be reviewed, and data abstraction will be performed using a standardized excel sheet designed for this study. Target data include baseline demographics and clinical data (e.g. age, parity, ethnicity, smoking, IVF pregnancy, history of gynecologic surgeries, type of twin pregnancy, current medical disorders, current obstetric complications, fetal anomalies, administration of antenatal steroids, Placental site, and twin-specific complications). Information from serial ultrasound reports including fetal growth and Doppler studies will be collected and data on fetal intervention will be abstracted. Peripartum data include node of delivery, Method of induction, CS indication, and type of cesarean incision. Clinical outcomes include postpartum hemorrhage, and perinatal death, admission to neonatal intensive care unit (NICU), neonatal need for respiratory support, neonatal intracranial hemorrhage, neonatal respiratory distress syndrome and neonatal hypoxic ischemic encephalopathy. Data will not include any identifiable information.

Prediction model will be created using baseline demographic and obstetric features of pregnancy and individual maternal and perinatal complications will be set as outcomes (dependent variables). A composite outcome of major maternal and neonatal outcomes will be created separately.

Inclusion Criteria:

1. Women with twin pregnancy who received their antenatal care and were delivered in the participating institute
2. Compliance to antenatal care visits

Exclusion Criteria:

1. Major fetal anomalies of one or both twins
2. Elective miscarriage
3. Authorization to use medical records was not provided by the patient

This study will include eligible women from at least 12 centers(anticipated) from different countries that present at least Europe, South America, Asia, and Africa.