Background: Increased ultrasound surveillance of twin pregnancies has become accepted practice due to the higher risk of complications. There is no current consensus however as to the method and frequency of ultrasound monitoring that constitutes optimal care. Objectives: To systematically review the effects of different types and frequency of ultrasound surveillance for women with a twin pregnancy on neonatal, fetal and maternal outcomes. Search methods: We searched Cochrane Pregnancy and Childbirth's Trials Register, ClinicalTrials.gov, the WHO International Clinical Trials Registry Platform (ICTRP) (all searched 11 August 2017), and reference lists of retrieved studies. Selection criteria: Randomised and quasi-randomised trials (including those published in abstract form) comparing the effects of described antenatal ultrasound surveillance regimens in twin pregnancies. Trials using a cluster-randomised design would have been eligible for inclusion in this review but none were identified. Trials using a cross-over design are not eligible for inclusion in this review. Different types and frequencies of ultrasound testing (for fetal surveillance and detection of specific problems) compared with each other and also compared with no testing. For example, an intervention might comprise a specific approach to ultrasound examination with dedicated components to detect twin-specific pathology. Different interventions could also include a specific type of surveillance at different intervals or different combinations at the same intervals. In this review we only found one study looking at fetal growth (biometry) and Doppler ultrasounds at 25, 30 and 35 weeks' gestation versus fetal growth alone. Data collection and analysis: Two review authors independently assessed trials for inclusion and quality, and extracted data. We checked data for accuracy. Main results: We included one trial of 526 women with a twin pregnancy of two viable twins, with no known morphological abnormality, in this review. The trial compared women receiving fetal growth and Doppler ultrasounds at 25, 30 and 35 weeks' gestation to fetal growth alone. We judged the included study to be at low risk of bias however the risk of performance and detection bias were unclear. The primary outcome was the perinatal mortality rate (after randomisation), for which there was no evidence of a clear difference between the fetal growth + Doppler and the fetal growth alone groups (risk ratio (RR) 0.88, 95% confidence interval (CI) 0.32 to 2.41, low-quality evidence) with similar rates in both groups (seven events in the Dopper + fetal growth group and eight in the fetal growth alone group). No clear differences were seen between the two regimens for the other outcomes in this review: stillbirth (RR 0.67, 95% CI 0.11 to 3.99), neonatal death (RR 1.01, 95% CI 0.29 to 3.46, low-quality evidence), gestational age at birth (weeks) (mean difference 0.10, 95% CI -0.39 to 0.59, moderate-quality evidence), infant requiring ventilation (RR 0.86, 95% CI 0.59 to 1.25), admission to special care or intensive care units (RR 0.96, 95% CI 0.88 to 1.05), caesarean section (any) (RR 1.00, 95% CI 0.81 to 1.23, high-quality evidence), elective caesarean section (RR 1.06, 95% CI 0.77 to 1.47), emergency caesarean section (RR 0.93, 95% CI 0.66 to 1.32), induction of labour (RR 1.10, 95% CI 0.80 to 1.50, moderate-quality evidence) or antenatal hospital admission (RR 0.96, 95% CI 0.80 to 1.15, high-quality evidence). The number of preterm births before 28 weeks' gestation was not reported in the included study. For the mortality-related outcomes, event numbers were small. The included study did not report the majority of our maternal and infant secondary outcomes. Infant outcomes not reported included fetal acidosis, Apgar scores less than 7 at five minutes and preterm birth before 37 and 34 weeks' gestation. The maternal outcomes; length of antenatal hospital stay, timely diagnosis of significant complications, rate of preterm, prelabour rupture of membranes and women's level of satisfaction with their care were not reported. The study did not classify twin pregnancies according to their chorionicity. An awareness of the chorionicity may have improved applicability of this data set. We downgraded outcomes assessed using GRADE for imprecision of effect estimates. Authors' conclusions: This review is based on one small study which was underpowered for detection of rare outcomes such as perinatal mortality, stillbirth and neonatal death. There is insufficient evidence from randomised controlled trials to inform best practice for fetal ultrasound surveillance regimens when caring for women with a twin pregnancy. More studies are needed to evaluate the effects of currently used ultrasound surveillance regimens in twin pregnancies. Future research could report on the important maternal and infant outcomes as listed in this review.