Risk of fetal loss following amniocentesis or chorionic villous sampling in twin pregnancies: a systematic review and meta-analysis

Objectives: To assess the rate of fetal loss following amniocentesis and chorionic villous sampling (CVS) in twin pregnancies. Methods: Medline, Embase and Cochrane databases were searched for studies reporting procedure-related complications following amniocentesis or chorionic villous sampling in twin pregnancies. The primary outcome was the rate of procedure-related fetal loss. The secondary outcomes were fetal loss occurring before the 24th week of gestation and fetal loss occurring within 4 weeks after the procedure. Head-to-head meta-analyses were used to directly compare, for each outcome: (a) women undergoing amniocentesis versus women not undergoing amniocentesis, and (b) women undergoing CVS versus women not undergoing CVS, and to compute pooled risk differences between women exposed and not exposed to each invasive procedure. Additionally, meta-analyses of proportions were used to estimate the pooled rates of each of the three outcomes among women undergoing amniocentesis or CVS, and among controls, and head-to-head meta-analyses to directly compare, for each outcome: women undergoing amniocentesis versus women not undergoing amniocentesis; women undergoing CVS versus women not undergoing CVS. Results: Sixteen studies (3419 twin pregnancies undergoing and 2517 twin pregnancies not invasive procedures) were included. In twin pregnancies undergoing amniocentesis, head-to-head meta-analyses directly comparing women undergoing amniocentesis versus women not undergoing amniocentesis found a higher risk for overall fetal loss (OR 1.46, p=0.04; RD 0.013, p=0.04), while there was no difference when focusing on either fetal loss before 24 weeks of gestation (OR 1.59, p=0.06; RD 0.010, p=0.11) or fetal loss within 4 weeks from the procedure (OR 1.38, p=0.3; RD 0.003, p=0.8). In twin pregnancies undergoing CVS, head-to-head meta-analyses directly comparing women undergoing CVS versus women not undergoing CVS found no significant difference either when investigating overall fetal loss (OR 1.61, p=0.5; RD 0.003, p=0.8) or fetal loss before 24 weeks of gestation (OR 1.61, p=0.5; RD 0.003, p=0.8). Conclusion: The risk of fetal loss following amniocentesis and CVS in twins is lower than the one previously reported, and the rate of fetal loss before 24 weeks of gestation or within 4 weeks from the procedure did not differ from the background risk of a twin pregnancy not undergoing invasive prenatal testing. These data can guide prenatal counselling for twin pregnancies undergoing invasive procedures.