Abstract

Objectives The primary aim of this systematic review was to explore the strength of association between birth‐weight (BW) discordance and perinatal mortality in twin pregnancy. The secondary aim was to ascertain the contribution of gestational age and growth restriction in predicting mortality in growth‐discordant twins. Methods MEDLINE, EMBASE, CINAHL and ClinicalTrials.gov databases were searched. Only studies reporting on the risk of mortality in twin pregnancies affected compared with those not affected by BW discordance were included. The primary outcomes explored were incidence of intrauterine death (IUD), neonatal death (NND) and perinatal death. Outcome was assessed separately for monochorionic (MC) and dichorionic (DC) twin pregnancies. Analyses were stratified according to BW discordance cut‐off (≥ 15%, ≥ 20%, ≥ 25% and ≥ 30%) and selected gestational characteristics, including incidence of IUD or NND before and after 34 weeks' gestation, presence of at least one small‐for‐gestational age (SGA) fetus in the twin pair and both twins being appropriate‐for‐gestational age. Risk of mortality in the larger vs smaller twin was also assessed. Meta‐analyses using individual data random‐effects logistic regression and meta‐analyses of proportion were used to analyze the data. Results Twenty‐two studies (10 877 twin pregnancies) were included in the analysis. In DC pregnancies, a higher risk of IUD, but not of NND, was observed in twins with BW discordance ≥ 15% (odds ratio (OR) 9.8, 95% CI, 3.9–29.4), ≥ 20% (OR 7.0, 95% CI, 4.15–11.8), ≥ 25% (OR 17.4, 95% CI, 8.3–36.7) and ≥ 30% (OR 22.9, 95% CI, 10.2–51.6) compared with those without weight discordance. For each cut‐off of BW discordance explored in DC pregnancies, the smaller twin was at higher risk of mortality compared with the larger one. In MC twin pregnancies, excluding cases affected by twin–twin transfusion syndrome, twins with BW discordance ≥ 20% (OR 2.8, 95% CI, 1.3–5.8) or ≥ 25% (OR 3.2, 95% CI, 1.5–6.7) were at higher risk of IUD, compared with controls. MC pregnancies with ≥ 25% weight discordance were also at increased risk of NND (OR 4.66, 95% CI, 1.8–12.4) compared with those with concordant weight. The risk of IUD was higher when considering discordant pregnancies involving at least one SGA fetus. The overall risk of mortality in MC pregnancies was similar between the smaller and larger twin, except in those with BW discordance ≥ 20%. Conclusion DC and MC twin pregnancies discordant for fetal growth are at higher risk of IUD but not of NND compared with pregnancies with concordant BW. The risk of IUD in BW‐discordant DC and MC twins is higher when at least one fetus is SGA.