Our aims were to assess the incidence of birthweight discordance of twins, to explore risk factors and its impact on perinatal mortality, and to quantify the risks at different severity of birthweight discordance in Japan, by using a nationwide obstetric database. There were 10,828 pairs of twins, born from 2001 to 2005 recorded in the database of the national Perinatal Health Care Project, fully enrolled. The overall incidence of birthweight discordance was 47.34%. The incidence of mild, severe and extremely severe discordance was 19.26%, 10.21% and 17.87%, respectively. The incidence of birthweight discordance in Japan is much higher than that in other countries, particularly at higher severity level. By linear regression model, our study added independent factors of primiparity (p < .001), sex composition (p < .001), chorionicity (p < .001), gestational age (p < .001), and delivery mode (p < .001) in determining birthweight discordance percentage. Maternal age and application of assisted reproduction technologies (ART) didn't significantly influence the birthweight discordance. The birthweight discordance is closely associated with gestational age and affected discharge mortality. From 25% of birthweight discordance, risk to discharge mortality tended to significantly increase, suggesting it should be added as a reference for clinical practices.

To evaluate whether growth discordance is an independent risk factor in the neonatal outcome of the smaller twin, all medical records of twin pregnancies delivered between 26 and 41 weeks during a 5-year period (January 2004–December 2008) were reviewed. Among the 49 selected twins, weight discordance was 15–20% in 7 infants, 21–30% in 16 infants, 31–40% in 16 infants and > 40% in 10 infants. No significant differences between the four groups were found with regards to obstetric complications and neonatal disease. Occurrence of birthweight below the 10th percentile and rate of admission to the neonatal intensive care unit significantly increased as intra-pair birthweight difference increased (p = .03). The > 40% discordant group had a significantly lower gestational age (p = .03), lower birthweight (p = .007) and a significantly higher mortality rate (4/10 versus 3/39 p = .04) in comparison with the other discordant groups. Multiple logistic regression analysis showed that birthweight was the single independent and consistent factor associated with elevated risks of mortality. For every 250 g increase in birthweight, the risk for mortality decreased by about 84% [RR 0.16(CI 0.00-0.70)]. Gestational age was the most reliable predictor for major neonatal complications. For every 1-week increase in gestational age a significant decreased risk for all outcomes was found. Discordance alone should not be considered as a predictor for adverse neonatal outcome. Neonatal outcome in discordant twins appears to be related to gestational age and birthweight rather than to the degree of discordance.