Causes of Delivery and Outcomes of Very Preterm Twins Stratified to Zygosity

The increasing rates of preterm birth among twins implicate that solid data on associated risks and outcomes are required. Assessment of zygosity is often based on clinical criteria (evaluation of placenta; same gender, birth weight discordance as surrogate criteria for monochorionic/monozygotic twins). The aim of this study was to compare clinical versus genetic assessment of zygosity and to compare causes of preterm delivery as well as outcome data of very-low-birth-weight (VLBW; birth weight <1,500 g) twins stratified to zygosity. In a multicenter study, we selected n = 176 sets of same gender twins and determined zygosity genetically. In a subgroup of 123 sets of twins, the attending physicians at the study centers were asked to document the parameter ‘zygosity’ (monozygotic/dizygotic) on the basis of their clinical judgment. Concordance between genetic and clinical assessment was 62.7% for monozygotic twins and 88.9% for dizygotic twins, respectively. Outcome parameters (death, BPD, ROP, NEC, IVH) were comparable in both groups. Genetically dizygotic twins were significantly more often born due to intrauterine infection (33% vs. 20% in monozygotic twins, p < .01) and antenatal antibiotics were more frequently given to mothers of dizygotic twins (62% vs. 47% in monozygotic twins, p < .01). Obstetric complications such as twin-twin-transfusion-syndrome were only seen in monozygotic twins as expected. The unexpected increase of antenatal antibiotic treatment and birth due to intrauterine infection in dizygotic twins should be confirmed in additional VLBW twin-cohorts.

Preterm twin birth rates have been increasing over the years (Hartley & Hitti, 2010). They are associated with significant long-term morbidity and mortality. Monozygotic twins are traditionally thought to suffer more complications mainly due to the risk for twin-twin-transfusionsyndrome (TTTS). Assessment of zygosity is often based on clinical criteria (evaluation of placenta; same gender, birth weight discordance as surrogate criteria for monochorionic/monozygotic twins). The rising rates of preterm birth among twins implicate that solid data on associated risks and outcome are required. Papiernik et al. (2010, MOSAIC-Study Group) compared very preterm twins (n = 1,254) and singletons (n = 3,585) with regard to prenatal and postnatal complications. In sin-gletons (<32 weeks of gestation), the investigators noted a higher rate of pregnancy complications (hypertension, hemorrhage) and growth restriction. While twins born between 24 and 27 weeks of gestation had a higher morbidity and mortality than singletons, twins between 28 and 31 weeks of gestation had a better outcome than singletons. In their cohort, same sex twins with a discordant weight had RECEIVED 13 September 2011; ACCEPTED 14 November 2011. ADDRESS FOR CORRESPONDENCE: Dr. Juliane Spiegler, Department of Pediatrics, University Hospital of Schleswig-Holstein, Ratzeburger Allee 160, 23538 Luebeck, Germany. E-mail: uni@drspiegler.de significantly higher mortality rates and an increased risk for IVH (intraventricular hemorrhage) grade III and IV. Their data did not include genetic assessment of zygosity.
The objective of this study was to compare the validity of clinical and genetic assessment of zygosity. Our second objective was to compare outcome data of very-low-birthweight (VLBW, birth weight <1,500 g) twins stratified to zygosity to test the hypothesis that monozygotic VLBW twins have more postnatal complications compared to dizygotic twins and VLBW singletons.
In different gender twins prenatal risk factors are identical since they share the same womb; they are born due to the same reasons and are treated in the same neonatal intensive care unit (NICU). Therefore, this group is very interesting to study gender effects. As a third objective, we compared dizygotic different gender VLBW twins to study the influence of gender in dizygotic twins.

Subjects
Data from VLBW infants born between 24+0 and 36+6 weeks of gestation and a birth weight of less than 1,500 g were collected from 28 centers in Germany. Between 2003 and 2009, a total of 3,708 VLBW infants with DNA samples were included in our study of influences of genetic polymorphisms on long-term morbidity and mortality. DNA of all infants was isolated from buccal swabs or umbilical cord tissue. Proper assessment of clinical data was ensured by onsite-monitoring. In a subgroup of twins enrolled between 2003 and 2007 (123 sets of twins), the attending physicians at the local study centers were asked to document the parameter 'zygosity' (monozygotic/dizygotic). Sets of twins classified as 'unknown' were excluded. Within the cohort of 1,173 multiples (31% of the total cohort), we selected 176 twin pairs (352 infants) of the same gender and both with a birth weight below 1,500 g and determined zygosity genetically.

Definitions
Zygosity was determined as described by Wurmb-Schwark et al. (2004). Bronchopulmonary dysplasia (BPD) was defined as need for supplemental oxygen at 36 weeks postmenstrual age. Mechanical ventilation was defined as any kind of mechanical ventilation during hospital stay. IVH was per definition any bleeding into the cerebral germinal matrix or the ventricles. Sepsis was defined as any sepsis with clinical symptoms, laboratory findings, and positive blood culture during the stay in the hospital. Surgery for retinopathy of prematurity (ROP) was defined as any ROP treated with laser -or cryocoagulation. Surgery for NEC/FIP was defined as any surgery necessary for necrotizing enterocolitis or focal intestinal perforation, surgery for PDA was any surgical ligation for persistent ductus arteriosus.

Statistical Analyses
Data analysis was performed using SPSS 17.0 data analysis package (IBM). Hypotheses were evaluated with Fisher Exact test. Quantitative data were compared with t-test. A p value <.05 was considered statistically significant.

Ethics
The study was approved by the local committee on research in human subjects of the University of Luebeck, and the local committees of all participating centers.
Clinical data of same gender twins and a control group of 2,535 VLBW singletons are given in Tables 1 and 2. Singletons were used as a control group since they represent the 'overall risks of prematurity' in the participating centers.
As expected, TTTS was only noted in monozygotic twins. Antenatal antibiotics were applied more often to dizygotic twins as compared to monozygotic twins or singletons. With regard to causes of preterm delivery, suspected intrauterine infection was noted more often in dizygotic twins. Major outcome parameter did not differ between the two groups ( Table 2). Outcome parameter of different gender twins stratified to gender is given in Table 3. Regarding major outcome parameter, there were no significant gender differences.

Discussion
Preterm twin birth rates have been increasing over the last decades for several reasons. Maternal age is higher in developed countries and assisted reproduction is more often necessary at an advanced age. Regarding twin pregnancies, Prapas et al. (2006) showed that maternal age is a risk factor for having VLBW twins. Hartley and Hitti (2010) demonstrated that the rate of preterm twin births increased from 1980 to 2005. While the rate of preterm birth at a very low gestational age (24-31 weeks) remained stable, the rate of preterm twins born between 32 and 36 weeks increased along with increasing requirement for respiratory support. Preterm twins are a highly susceptible group in the NICUs and their special risks need to be identified more clearly (Spiegler et al., 2008). Note: Data are presented as mean and SD (±) or as total numbers and percentage. Note: Data are presented as mean and SD (±) or as total numbers and percentage; BPD -bronchopulmonary dysplasia, Sepsis -blood-culture-proven sepsis, IVH -intraventricular hemorrhage, NEC -necrotizing enterocolitis, ROP -retinopathy of prematurity. Note: Data are presented as mean and SD (±) or as total numbers and percentage; BPD -bronchopulmonary dysplasia, Sepsis -blood-cultureproven sepsis, IVH -intraventricular hemorrhage, NEC -necrotizing enterocolitis, ROP -retinopathy of prematurity.
All data in our study are limited by the relatively small number in each group. As in Sudanese twins (Elshibly & Schmalisch, 2010) and in other studies, girls weighed less than boys in our cohort; however, this did not reach statistical significance. In their study, twins were more prone to be growth restricted than singletons. In our study, birth weight of all groups of twins is very similar to singletons. However, the Sudanese twins were significantly less premature (average gestational age 37 weeks, compared to 28 weeks in our study). Interestingly, none of the major outcome parameters of different gender twins reached significance even though strong gender effects have been shown in large groups of preterm babies (Tyson et al., 2008).
In our cohort, causes of preterm delivery were similar between mono-and dizygotic twins. Monozygotic twins received prenatal steroids more frequently (95% in monozygotic twins, 90% in dizygotic twins, and 87% in singletons). This trend did not reach statistical significance; however, this trend might reflect that monozygotic twins are supposed to be more at risk than dizygotic twins or singletons. Myles et al. (1997) reported that twin pregnancies with premature rupture of membranes (PROM) had a shorter latency until birth but were less likely to suffer from amniotic infection. Mazor et al. (1996) demonstrated that the prevalence of intraamniotic infection was similar in twin and singleton pregnancies with preterm labor. In our study, dizygotic twins were more prone to receive antenatal antibiotics and birth was more likely due to suspected intrauterine infection with preterm labor and rupture of membranes. PROM is seen more often in twin gestations. Our cohort was not monitored for PROM.
However, even a higher rate of PROM does not explain why dizygotic twins are more prone to suspected infection than monozygotic twins. Dizygotic same gender twins should have a similar risk of infection as different gender twins. In our cohort, we saw a similar high rate of antenatal antibiotics in different gender twins compared to dizygotic same gender twins (60% vs. 62%) but less suspected intrauterine infection (22% vs. 33%). If premature labor does not respond to tocolytics and antenatal antibiotics, intrauterine infection is a frequent cause of preterm delivery. Despite receiving more tocolytics (75%), premature labor was the most common reason for preterm birth in different gender twins (59% vs. monozygotic twins 44%, same gender dizygotic twins 46%, singletons 30%). Larger cohorts are needed to confirm whether or not dizygotic twins are more prone to infection than monozygotic twins.
Cesarean section is the standard delivery mode in most German delivery wards for VLBW infants, while the MO-SAIC group reported rates of spontaneous delivery up to 52%. There is an ongoing discussion whether or not a cesarean section improves outcome in VLBW infants (Paul et al., 2002;Riskin et al., 2008). Papiernik et al. (2010) showed in a large cohort (MO-SAIC group) that twins at 24-27 weeks of gestation had an increased neonatal mortality and higher rates of intracranial hemorrhage if they were of the same gender and had discordant birth weights. This suggests an influence of zygosity (and suspected TTTS) on the observed morbidity. However, no genetic determination of zygosity was done. In our cohort, clinical assessment of zygosity alone revealed inconsistent results in 10-30%. Therefore, clinical assessment of zygosity on its own may not lead to proper counseling with regard to outcome of twins. After correcting for zygosity, we observed no significant differences in major outcome parameters (including IVH) stratified to zygosity (Table 2). Larger cohorts are necessary to confirm our observation, given the fact that our data contradict previous data (Papiernik et al., 2010).
In 1999, Mizrahi et al. reported that twins do better than comparable singletons regarding peripartal complications. Kirkby et al. (2010) compared outcome of VLBW twins and higher multiples to VLBW singletons and discovered no significant differences. The MOSAIC-Group showed differences in same gender twins between 24 and 27 weeks of gestation with a discordant birth weight, but otherwise similar outcomes of twins and singletons. Together with our data, there are now four cohorts that show comparable outcome data for VLBW twins and singletons in North America, different areas in Europe (MOSAIC-Group), Israel, and Germany (German Neonatal Network). Our data are useful and reassuring for counseling parents of VLBW twins. The unexpected observation of a higher rate of intrauterine infection in dizygotic VLBW twins should be confirmed in a larger cohort.

Conclusion
Clinical assessment of zygosity in preterm twins on its own was unreliable and may be a significant confounder of data. Therefore, genetic assessment of zygosity has the potential to optimize data quality in VLBW twins. Consequently, prognostic counseling of parents on mortality and longterm morbidity of preterm twins may be improved. If data of VLBW-twin pregnancies are analyzed for scientific purposes, genetic determination of zygosity is highly recommended. Major outcome parameters show no differences between monozygotic twins, dizygotic twins, and singletons. Regarding prenatal complications, we found a higher risk for suspected intrauterine infection and subsequent premature delivery in dizygotic twins that needs to be confirmed in larger cohorts.