Objective
We sought to estimate the optimal time to deliver uncomplicated monochorionic-diamnionic (MCDA) twins.
Study Design
Data were retrospectively obtained from twin pregnancies from 2000 through 2009. The gestational week–specific prospective perinatal mortality risk was calculated. A cohort of MCDA twins with nonindicated deliveries was analyzed separately. Neonatal outcomes and costs were compared between MCDA twins with nonindicated deliveries born at specific weeks of gestation, and those born the subsequent week.
Results
There were 5894 dichorionic-diamnionic twins and 1704 MCDA twins. After 28 weeks, the gestational week–specific prospective risk of perinatal mortality did not differ between groups. There were 948 MCDA twins with nonindicated deliveries. Until 37 weeks, the risk of severe neonatal morbidity, perinatal mortality, and hospital costs were greater for fetuses delivered compared to fetuses born in a subsequent week.
Conclusion
To optimize neonatal outcome and decrease hospital costs, MCDA twins should not be delivered <37 weeks unless medically indicated.
The number of multiple pregnancies is steadily increasing. From 1980 through 2006 in the United States, the number of infants born from twin gestations has increased by >60%. Currently, multiple gestations comprise approximately 4% of all live births; however, they contribute almost 20% to overall perinatal morbidity and mortality rates. Monochorionicity has been associated with increased adverse perinatal outcomes when compared with dichorionicity. This is in part due to higher incidence of intrauterine growth restriction, abnormal placental vascular anastomoses, fetal anomalies, and stillbirth. A higher rate of prematurity in monochorionic twins also contributes to their increased rate of adverse perinatal outcomes. Many obstetricians now advocate iatrogenic preterm delivery of monochorionic twins to avoid potential perinatal complications such as stillbirth. Notably, the rate of preterm births in twin pregnancies in North America has increased from approximately 30% in the 1970s to >50% in the late 1990s. Monochorionic twins are often delivered in clinical practice at significantly earlier gestational ages than dichorionic twins.
For Editors’ Commentary, see Table of Contents
It is well known that patients with monochorionic twins are more likely to develop complications including twin-twin transfusion syndrome, fetal growth abnormalities, and severe preeclampsia–all of which warrant preterm delivery. However, the optimal timing of delivery for monochorionic twin pregnancies that do not have medical indications for delivery remains a subject of debate. Some experts argue that the risk of fetal death in monochorionic twin pregnancies is high enough to recommend elective preterm delivery as early as 32 weeks’ gestation. Others disagree and recommend delivery >36 weeks.
It has become increasingly clear that there are significant adverse perinatal outcomes associated with “late preterm birth.” Neonates born at 34-37 weeks have an increased risk of serious neonatal morbidity when compared with term infants. With this in mind, the optimal timing of delivery of monochorionic twins requires reexamination. While prior studies have examined the timing of delivery of monochorionic twins, study populations were small and results were conflicting. The recently published Eunice Kennedy Shriver National Institute of Child Health and Human Development Workshop White Paper on pregnancy and pregnancy outcomes also addresses the timing of delivery and emphasizes the importance of reducing iatrogenic preterm birth. However, there are no nationally established recommendations regarding timing of delivery for uncomplicated monochorionic twins.
The objective of this study was to estimate the optimal timing of delivery in a large cohort of monochorionic-diamnionic (MCDA) twins that did not have any medical indications for delivery. We hypothesized that the morbidity and mortality rates of MCDA twins are similar to those of diamnionic twins, that the risk of prematurity associated with elective delivery <37 weeks’ gestation is greater than the risk of stillbirth, and therefore MCDA twins without medical indications for delivery do not warrant elective preterm delivery.
Materials and Methods
We performed a retrospective cohort study and institutional review board approval was obtained. Electronic data were obtained on patients with twin deliveries from January 2000 through December 2009 at 18 hospitals within the Intermountain Health Care medical system. All patients with monoamnionic twins and those with unknown chorionicity were excluded. Chorionicity was determined using prenatal ultrasound assessment of the presence of a single vs fused, or separate placentas; thickness of the dividing membrane; the presence or absence of a “twin peak” sign on first- or second-trimester ultrasound; concordance or discordance of fetal gender; and postpartum pathological examination of placentas and membranes. Gestational age was determined by patients’ last menstrual periods and confirmed by ultrasonography. In cases where the menstrual dating was unknown or was discordant with first-trimester or early second-trimester ultrasound measurements, the ultrasonographically determined gestational age was used.
Data from neonatal outcomes were collected and analyzed. We defined perinatal death as stillbirth (fetal death prior to delivery) or neonatal death (death by 28 days in a liveborn infant). For adverse perinatal events analyses, we defined severe adverse perinatal events as stillbirth, neonatal death, bronchopulmonary dysplasia, grade >3 intraventricular hemorrhage, necrotizing enterocolitis, or sepsis. Survival-to-discharge analyses and curves were generated for monochorionic and dichorionic pregnancies. Cox proportional hazards regression method with adjusted covariates was performed to assess the degree of difference between the 2 groups. Gestational age–specific prospective risks of perinatal mortality were calculated using the concept of “fetuses at risk” as previously described. For this calculation the number of stillbirths and neonatal deaths during a given week “W” was divided by the number of fetuses remaining in utero at the beginning of week “W” (fetuses at risk). This calculation was done at weekly increments and compared between MCDA and dichorionic-diamniotic (DCDA) twins.
To determine the optimal timing of delivery we examined 2 approaches to calculate associated risks, namely, gestational age–specific prospective risk as described above, and “subsequent risk.” Hospital charges that were also obtained directly from the electronic medical record were examined separately. For these analyses, we utilized only deliveries without medical indications, which we defined as cases without preeclampsia, diabetes, autoimmune disease, twin-twin transfusion syndrome (diagnosed according to the criteria of Quintero et al ), intrauterine growth restriction (estimated fetal weight <10%), fetal anomalies, and/or chorioamnionitis. Data from fetal heart rate tracings were not available for analysis.
For subsequent risk analyses we compared neonatal risk for MCDA twins in a given week with subsequent risk. Neonatal risk was defined as: [number of events (excluding fetal deaths) during week “W”]/[live births during week “W”]. Subsequent risk was defined as: [number of events after week “W” + stillbirths during week “W”]/[deliveries after week “W” + stillbirths during week “W”]. In this comparison, all twins could be unambiguously assigned to a single cell in a 2-way classification and risk differences (events associated with current or subsequent weeks) tested by standard χ 2 testing of 2 × 2 tables. Subsequent risk provides a somewhat different perspective than the gestational age–specific prospective risk described above, in that it calculates the risk of continuing a pregnancy past week “W” rather than the risk to the pregnancy specifically at week “W.” Furthermore, this rate can be directly compared to the risk experienced by neonates delivered week “W.” Neonatal hospital charges for MCDA twins with nonindicated deliveries were calculated among live births during a given week and compared to subsequent births at weekly increments.
Results
A total of 4272 women with twin pregnancies delivered during the study period. In all, 64 patients with monoamnionic twins and 409 patients in whom the chorionicity was not determined were excluded. Overall 3799 women with twin gestations and their 7598 neonates were included in the analysis. There were 852 women with MCDA twins (1704 infants) and 2947 women with DCDA twins (5894 infants). Of all women, 2452 had deliveries that were not medically indicated and delivered 4904 neonates. Among these nonindicated deliveries, there were 3956 (80.7%) infants from DCDA pregnancies and 948 (19.3%) from MCDA pregnancies ( Figure 1 ) .
