Perinatal Mortality Rates and Risk Factors for Mortality among Zygotic Twins and Singletons in Japan, 1995-2008

In Japan, the fetal death rate (FDR; defined as death after GA of ≥22 weeks) decreased significantly between 1980/81 and 1998 [3]. During this period, FDRs for monozygotic (MZ) and dizygotic (DZ) twins decreased from 73 to 32 per 1000 twin deliveries and from 33 to 10 per 1000 twin deliveries, respectively [3]. However, there is no information on early neonatal deaths (ENDs) and PMRs for zygotic twins and singletons in Japan.

shows the yearly change in FDR, ENDR, and PMR for MZ and DZ twins between 1995 and 2008. For MZ and DZ twins, each death rate significantly decreased every year during the study period. By 2008, each mortality rate declined to approximately half of the 1995 value for MZ twins, but this decrease was to a third or a quarter of the initial value for DZ twins. FDRs and PMRs were also significantly higher for MZ twins than DZ twins in each year. However, ENDRs were only significantly higher for MZ twins than for DZ twins in eight years (1995, 1996, 1999, 2000, 2003, 2005, 2007, and 2008). Table 2 shows the yearly changes in FDR, ENDR, and PMR for singletons during the study period. Each mortality rate significantly decreased every year during this period. By 2008, each mortality rate declined to approximately half of the 1995 value. Each year, the mortality rates were significantly higher for MZ twins and DZ twins than for singletons, and all mortality rates decreased every year. Table 3 shows FDRs, ENDRs, and PMRs for zygotic twins by MA group during the study period. FDR for MZ twins was the highest for MA of ≥ 40 years, followed by MA of <20 years and was lowest for MA of 35-39 years where the rate was significantly lower than MA of 20-24 years. The highest FDR for DZ twins was at MA of <20 years and the lowest was at MA 30-34 years where the rate was significantly lower than the rates for MAs of 20-24 years, 25-29 years, and 35-39 years. FDR was significantly higher for MZ twins than for DZ twins in each MA group, except when MA was <20 years. As for ENDRs, the highest rates were the youngest MA groups for both zygotic twins. ≥ 40 years. The highest PMRs were the youngest MA groups, and the lowest rates were at MA of 30-34 years for both zygotic twins. PMR was significantly higher for MZ twins than for DZ twins in each MA group, except for the MA group of <20 years. Table 4 shows FDRs, ENDRs, and PMRs for singletons according to MA during the study period. Each mortality rate was the highest at MA of ≥ 40 years and was the lowest at MA of 25-29 years; each mortality rate was significantly lower at MA of 25-29 years compared with those at other MAs.        Table 5 shows FDRs, ENDRs, and PMRs for zygotic twins by GA during the study period. FDR for MZ twins was 665 per 1000 twin deliveries at GA of <24 weeks but decreased with increasing GA to 37 weeks (5.7). However, FDR at 37 weeks' GA was significantly lower than that at <36 weeks and ≥ 40 weeks. FDR for DZ twins was 419 at GA of <24 weeks and decreased with increasing GA to 37 weeks (1.7), when FDR was significantly lower than that for other GA groups. ENDR for MZ twins was 581 for GA of <24 weeks and decreased with increasing GA to 37 weeks (0.8) and was significantly lower at 37 weeks than at either <36 weeks or at 38-39 weeks. In DZ twins, ENDR was 333 at GA of <24 weeks and decreased with increasing GA up to 39 weeks (0.7) before increasing by ≥ 40 weeks (4.3). ENDR at 39 weeks' GA was significantly lower than at either <36 weeks or ≥ 40 weeks. PMR for MZ twins was 859 at GA of <24 weeks and decreased with increasing GA to 37 weeks (6.6). The rate for DZ twins was 612 for GA of <24 weeks and decreased with increasing GA to 37 weeks (3.0). PMR for DZ twins was significantly lower at GA of 37 weeks than in the other GA groups but was significantly higher for MZ twins than for DZ twins per GA group, except for GA of ≥ 39 weeks. Table 4 also shows FDRs, ENDRs, and PMRs for singletons according to GA group during the study period. All rates were significantly lower at GA of ≥ 40 weeks compared with those at other GAs. Figure 2 shows PMR by GA for MZ twins, DZ twins, and singletons during the study period. PMRs were higher for singletons than for either the MZ or DZ twin cohorts between GA of <24 weeks and 32-35 weeks, but these PMRs were reversed at GA of 36-39 weeks and at GA of ≥ 40 weeks. Table 6 shows the preterm birth rates for MZ twins, DZ twins, and singletons during the study period. The preterm birth rate included all births with GA of <37 weeks and were calculated after excluding fetuses that had been delivered at GA of <22 weeks. For MZ twins, the rate was 44% in 1995 and gradually increased to 62% in 2008. For DZ twins, the corresponding rates were 39% and 55%. The rates for singletons were 4.6% in 1999 and 4.9% in 2008. The recent increase in preterm birth was associated with reduced PMRs for MZ and DZ twins.

Rate of preterm birth
Among the perinatal deaths during the study period, the preterm birth rate was 89% (1968.1/2216.2) for MZ twins and 83% (1020.5/1233) for DZ twins (data in Table 5).

Discussion
PMR was 6-fold higher for twins than for singletons in Japan between 1980 and 1991 [16]. In the present study, we showed that the relative risk of PMR for MZ twins versus singletons was 7-fold (45.6/6.46) higher than that of singletons in 1995 and that this risk decreased to 5.9-fold (23.5/4.01) in 2008; these relative risks were 3.8fold (24.7/6.46) and 1.9-fold (7.6/4.01) for DZ twins versus singletons in 1995 and 2008, respectively. The relative risk of PMR increased 2-to 3-fold between MZ and DZ twins during the study period. PMR was markedly improved for DZ twins than for MZ twins and singletons. As for FDRs, declines to approximately 1/4-1/3 for DZ twins and to 1/2 for both MZ twins and singletons were seen during the study period. Imaizumi [3] estimated FDRs for zygotic twins from 1980-1981 to 1998. However, ENDRs and PMRs for zygotic twins were estimated for the first time in the present study.  Loos et al. [10] reported that the stillbirth rate was significantly higher for MZ monochorionic (MC) twins than for DZ twins in Belgium. Unfortunately, the data available to us did not include details of twin chorionicity, which precludes direct comparison. Glinianaia et al. [13] also reported that MC twins have higher stillbirth rates compared with MZ dichorionic (DC) twins in England for the period between 1998 and 2007.
The increased risk of stillbirth in MC twins than DC twins has been primarily attributed to twin-twin transfusion syndrome (TTTS) [11-13, 17,18]. In addition, Morikawa et al. [19] reported that Japanese women with MC, diamniotic twins were 2.2-fold more likely to experience stillbirth than those who had DC, diamniotic twins between 2005 and 2008. Imaizumi and Hayakawa [20] also reported that 14% of stillbirths in MZ twins were attributed to TTTS and that 4% of stillbirths were due to birth defects, whereas the corresponding vales in DZ twins were 0% and 3%, respectively, between 1995 and 2008. In the present study, PMRs for MZ and DZ twins significantly decreased between 1995 and 2008, and the rate was significantly higher for MZ twins than for DZ twins every year. The higher PMR for MZ twins than DZ twins could be attributed to the higher rates of TTTS and birth defects.
In a comprehensive literature review, Mercuro et al. [21] reported that preterm birth and low BW contributed to an increase in cardiovascular risk in later life. Although preterm birth rates increased by 18% for MZ twins and by 16% for DZ twins from 1995 to 2008, PMR decreased for both MZ and DZ twins. As for singletons, preterm birth rates increased by only 0.3% from 1999 to 2008. It is therefore plausible that the higher rate of prematurity in twins increases the risk of late-life complications, such as cardiovascular risk, in this group than in singletons.