Optimal timing of fetal reduction from twins to singleton: earlier the better or later the better?

To determine the rate of pregnancy complications and adverse obstetric and neonatal outcomes of twin pregnancies that were reduced to singleton at an early compared with a later gestational age.

This was a historical cohort study of dichorionic diamniotic twin pregnancies that underwent fetal reduction to singletons in a single tertiary referral center between January 2005 and February 2017. The study population was divided into two groups according to gestational age at fetal reduction: those performed at 11-14 weeks' gestation, mainly at the patient's request or as a result of a complicated medical or obstetric history; and selective reductions performed at 15-23 weeks for structural or genetic anomalies. The main outcome measures compared between pregnancies that underwent early reduction and those that underwent late reduction included rates of pregnancy complications, pregnancy loss, preterm delivery and adverse neonatal outcome.

In total, 248 dichorionic diamniotic twin pregnancies were included, of which 172 underwent early reduction and 76 underwent late reduction. Although gestational age at delivery was not significantly different between the late- and early-reduction groups (38 weeks, (interquartile range (IQR), 36-40 weeks) vs 39 weeks (IQR, 38-40 weeks); P = 0.2), the rates of preterm delivery < 37 weeks (28.0% vs 14.0%; P = 0.01), < 34 weeks (12.0% vs 1.8%; P = 0.002) and < 32 weeks (8.0% vs 1.8%; P = 0.026) were significantly higher in pregnancies that underwent late reduction. Regression analysis revealed that late reduction of twins was an independent risk factor for preterm delivery, after adjustment for maternal age, parity, body mass index and the location of the reduced sac. Rates of early complications linked to the reduction procedure itself, such as infection, vaginal bleeding and leakage of fluids, were comparable between the groups (7.0% for early reduction vs 9.2% for late reduction; P = 0.53). There was no significant difference in the rate of pregnancy loss before 24 weeks (0.6% for early reduction vs 1.3% for late reduction; P = 0.52), and no cases of intrauterine fetal death at or after 24 weeks were documented. There was no significant difference in the prevalence of gestational diabetes mellitus, hypertensive disorders of pregnancy, preterm prelabor rupture of membranes or small-for-gestational age. The rates of respiratory distress syndrome (6.7% vs 0%; P = 0.002), need for mechanical ventilation (6.7% vs 0.6%; P = 0.01) and composite neonatal morbidity (defined as one or more of respiratory distress syndrome, sepsis, necrotizing enterocolitis, intraventricular hemorrhage, need for respiratory support or neonatal death) (10.7% vs 2.9%; P = 0.025) were higher in the late- than in the early-reduction group. Other neonatal outcomes were comparable between the groups.

Compared with late first-trimester reduction of twins, second-trimester reduction is associated with an increased rate of prematurity and adverse neonatal outcome, without increasing the rate of procedure-related complications. Technological advances in sonographic diagnosis and more frequent use of chorionic villus sampling have enabled earlier detection of fetal anatomic and chromosomal abnormalities. Therefore, efforts should be made to complete early fetal assessment to allow reduction during the first trimester. Copyright © 2020 ISUOG. Published by John Wiley & Sons Ltd.