Objective
The purpose of this study was to compare strategies for delivery timing of uncomplicated monochorionic diamniotic twin pregnancies.
Study Design
A decision tree compared 9 strategies that included scheduled delivery between 32 and 38 weeks’ gestation, with or without confirmation of fetal lung maturity. Outcomes in the model included fetal death, infant death, respiratory distress syndrome, mental retardation, and cerebral palsy.
Results
A scheduled delivery at 38 weeks’ gestation was the preferred strategy, which resulted in the highest quality adjusted life years under base-case assumptions. Decreased, but comparable, quality adjusted life years estimates resulted from scheduled deliveries at 36 and 37 weeks’ gestation, with or without amniocentesis. Sensitivity analyses demonstrated that the optimal gestational age for delivery was always ≥36 weeks’ gestation.
Conclusion
This decision analysis suggests that, for women with uncomplicated monochorionic twins, delivery between 36 and 38 weeks’ gestation is the preferred strategy for timing of delivery.
Monochorionic diamniotic twins account for 20% of twin gestations. Women with these pregnancies have an increased risk of adverse perinatal outcomes that include preterm delivery, fetal growth restriction, fetal anomalies, twin-twin transfusion syndrome, and stillbirth. Of particular concern, even with antenatal surveillance that indicates concordant twin growth, reassuring fetal testing, and an absence of features that are suggestive of twin-twin transfusion syndrome, an increased risk of fetal death remains that may be as high as 4.6% after 32 weeks’ gestation. Moreover, when 1 fetus dies, the co-twin assumes a substantial risk of death or neurologic injury, which likely is related to transient hemodynamic instability within their shared placental circulation.
Because of these risks, some providers have suggested that “delivering apparently uncomplicated monochorionic twins at 34-35 weeks’ gestation after administration of antenatal corticosteroids … is reasonable provided patients have been counseled about the risks and benefits.” Yet, although scheduled early delivery negates the chance of fetal death beyond a selected gestational age, it exposes neonates to the risk of potentially severe complications of prematurity. Optimal care would balance the risk of fetal death and neurologic injury against prematurity-related complications. Unfortunately, there is limited evidence to guide physicians in their selection of an optimal gestational age for delivery and no international consensus regarding delivery timing. The 2011 UK National Institute for Health and Clinical Excellence Guidelines on Antenatal Management of Multiple Gestations suggest that 36 weeks’ gestational age “does not appear to be associated with an increased risk of serious adverse outcomes” and that continuing these pregnancies past 38 weeks increases the risk of fetal death. Although there is no official guidance from the American College of Obstetricians and Gynecologists or the Society for Maternal Fetal Medicine, a consensus recommendation that stems from a joint National Institute of Child Health and Human Development/Society for Maternal Fetal Medicine workshop states that a range of 34–37 weeks’ gestation is acceptable. In an effort to better quantify the tradeoffs in outcomes regarding the timing of delivery, we performed a decision analysis with the goal of determining the ideal gestational age at which to deliver women with otherwise uncomplicated monochorionic diamniotic twins.
Materials and Methods
This study was considered exempt by the institutional review board at Northwestern University. A decision analytic model was designed with DATA software (version 3.5; TreeAge Software Inc, Williamstown, MA) that compared 9 different strategies for the timing of delivery in individuals with a concordantly grown monochorionic diamniotic twin gestation and no other complications of pregnancy (eg, preeclampsia, intertwin discordance, fetal growth restriction, twin-twin transfusion; Table 1 ). All patients were presumed to have reached 32 weeks’ gestation with 2 viable twins. Strategies 1, 2, 3, and 4 include delivery at 32, 33, 34, and 35 weeks’ gestation, respectively, after previous administration of antenatal corticosteroids for neonatal benefit. The remaining strategies do not involve routine administration of antenatal corticosteroids. Strategies 5, 7, and 9 involve expectant management until the target gestational age of 36, 37, and 38 weeks’ gestation, respectively. Strategies 6 and 8 involve expectant management until an amniocentesis test for fetal lung maturity at the indicated gestational age has been performed. For strategy 6, if fetal lung immaturity is detected, then amniocentesis is repeated 1 week later, and delivery is performed when either the fetal lung maturity is confirmed or at 38 weeks’ gestation, whichever occurs first. Strategy 8 entails delivery at 37 weeks’ gestation if fetal lung maturity is documented, otherwise delivery occurs at 38 weeks’ gestation. In all strategies, expectant management is abandoned if medical or obstetric indications arise that prompt delivery or if spontaneous labor occurs.
| Strategy | Gestational age at scheduled delivery |
|---|---|
| 1 | 32 wks after steroid administration |
| 2 | 33 wks after steroid administration |
| 3 | 34 wks after steroid administration |
| 4 | 35 wks after steroid administration |
| 5 | 36 wks |
| 6 | 36 wks pending amniocentesis |
| 7 | 37 wks |
| 8 | 37 wks pending amniocentesis |
| 9 | 38 wks |
The base case estimates were chosen on the basis of data within the published literature. For the purpose of the model, all pregnancies in which a fetal loss occurred were delivered shortly thereafter for the potential benefit of the co-twin. If the loss occurred at <34 weeks’ gestation, delivery was delayed only for the administration of antenatal corticosteroids. Estimates of the prospective risk of fetal loss were calculated on a per-pregnancy basis as the number of stillbirths per pregnancy after a given gestational age divided by the total number of ongoing pregnancies at the start of the time period. For pregnancies in which a fetal death occurred, the chance of fetal death or neurologic injury of the co-twin were taken from metaanalytic data. Among survivors with neurologic injury, the ratio of cerebral palsy to mental retardation was chosen based on a longitudinal study of neurologic outcomes among co-twin survivors. Based on large population-based twin studies, the risks of neurologic disability and infant death are also increased when a co-twin experiences an infant death; thus, these estimates were included in the model. To determine the proportion of monochorionic twins who, even if treated expectantly, would require delivery for medical or obstetric indications, population data from >500,000 twin pregnancies were used.
Respiratory distress syndrome (RDS) estimates were derived from a multinational observational study that detailed neonatal morbidities of >20,000 neonates who were not exposed routinely to in utero corticosteroids. The rate of RDS is not higher in twins than in singleton neonates at the gestational ages included in the model. Data from a study of almost 10,000 twins were used to determine RDS concordance or the likelihood that both vs only 1 of the twins had RDS at a specific gestational age. The frequency of RDS for twin pregnancies in the model was assumed at baseline not to be altered by the administration of antenatal corticosteroids, although sensitivity analyses took into account the possibility that the incidence of RDS in neonates whose mothers who had received antepartum steroids could be reduced by as much as 50%. Amniocentesis that was performed to assess fetal lung maturity involved the collection of only 1 twin’s amniotic fluid. The probability that the amniocentesis revealed a mature value at each gestational age has been described previously. The assumption was made that amniocentesis would not alter the risk of gross membrane rupture, labor onset, or fetal death.
This model altered timing but not route of delivery; thus, the effectiveness of each strategy was evaluated on the basis of differences in perinatal, but not maternal, outcomes. The life expectancy for each twin was assumed to be 75 years. The quality-adjusted life years (QALYs) for each strategy were based on this anticipated life expectancy and were discounted annually over a range of 0–3%. Adverse perinatal outcomes that were considered in the model were perinatal death, RDS, cerebral palsy, mental retardation, and infant death. Whenever a range of values for a given variable was found in the literature, sensitivity analysis was used to evaluate the effect that choosing different values within the range of had on the results.
Newborn quantitative measurements of the value to the decision-maker of the various outcomes of a decision (utilities) for each health outcome were determined on the basis of the literature. Applied QALYs for both cerebral palsy and mental retardation were based on utilities for moderate cerebral palsy and moderate mental retardation. Utilities that were associated with RDS were applied for shorter durations than utilities for chronic conditions such as mental retardation and cerebral palsy; the decrement in quality of life that was associated with mental retardation and cerebral palsy was considered from 1 year of life onward. The gestational age–specific probabilities of cerebral palsy, mental retardation, and infant death were derived from singleton data. The literature described similar incidences of these adverse outcomes for the gestational ages that we included in this model between singleton and twin pregnancies in which both twins are liveborn when stratified by gestational age at delivery. Dead fetuses did not accrue any QALYs, and those infants who died within the first year were presumed to have died at and accrued QALYs for 3.5 months of life.
Results
When the baseline assumptions that are listed in Tables 2-4 are used, strategy 9 (which is a scheduled delivery at 38 weeks’ gestation) results in the highest total QALYs ( Figure ). Scheduled deliveries at 36 or 37 weeks’ gestation, with or without incorporation of amniocentesis for fetal lung maturity, follows closely behind delivery at 38 weeks’ gestation. The incorporation of amniocentesis when delivering women with monochorionic diamniotic twins at <38 weeks’ gestation appears to offer a negligible improvement in outcome when compared with outright delivery at the same gestational age.
| Outcome | Utility a | Duration of effect |
|---|---|---|
| Perinatal death | 0 | Anticipated life expectancy |
| Infant death | 0 | From 14 weeks of life through the anticipated life expectancy |
| Infant respiratory distress syndrome | ||
| At 32 wks’ gestation | 0.87 (0.87–0.94) | 7 wks |
| At 33 wks’ gestation | 0.87 (0.87–0.94) | 6 wks |
| At 34 wks’ gestation | 0.87 (0.87–0.94) | 5 wks |
| At 35 wks’ gestation | 0.87 (0.87–0.94) | 4 wks |
| At 36 wks’ gestation | 0.87 (0.87–0.94) | 3 wks |
| At 37 wks’ gestation | 0.87 (0.87–0.94) | 2 wks |
| At 38 wks’ gestation | 0.87 (0.87–0.94) | 1 wks |
| Cerebral palsy | 0.76 (0.60–0.87) | From 1 year of life through the anticipated life expectancy |
| Mental retardation | 0.79 (0.59–0.84) | From 1 year of life through the anticipated life expectancy |
a Data are given as baseline value (range in sensitivity analysis).
| Outcome | P value or RR a |
|---|---|
| Respiratory distress syndrome rate | |
| At 32 wk (with steroids) | .30 b (.30, .41) |
| At 33 wk (with steroids) | RR, 1.0 |
| At 34 wk (with steroids) | RR, 0.43 |
| At 35 wk (with steroids) | RR, 0.21 |
| At 36 wk without steroids | RR, 0.22 |
| At 37 wk without steroids | RR, 0.07 |
| At 38 wk without steroids | RR, 0.0055 |
| Respiratory distress syndrome rate in twin pregnancy after steroids, if given at 32-34 wk | RR, 1 (0.5, 1) |
| Rate of co-twin death after in utero loss of 1 twin | .12 (.07, .18) |
| Rate of neurologic injury | |
| After in utero death of co-twin | .18 (.11, .26) |
| If infant death of co-twin | .147 (0, .147) |
| If neurologic injury of co-twin | .118 (0, .275) |
| Fetal lung maturity (+) at 36 wk | .817 |
| Respiratory distress syndrome at 36 wk | |
| With fetal lung maturity (+) | .006 |
| Without fetal lung maturity (−) | .15 |
| Fetal lung maturity (+) at 37 wk | |
| If first amniocentesis | .830 |
| If no fetal lung maturity (−) at 36 wk | .50 |
| Respiratory distress syndrome at 37 wk | |
| With fetal lung maturity (+) | .001 |
| With no fetal lung maturity (−) | .022 |
| Fetal lung maturity testing | |
| Sensitivity | .85 |
| Specificity | .84 |
| Rate of respiratory distress syndrome concordance | .67 |
| Cerebral palsy rate after birth | |
| At 32-33 wk | .0177 b (.01076, .02302) |
| At 34-36 wk | RR, 0.412 |
| At 37-38 wk | RR, 0.113 |
| Mental retardation rate after birth | |
| At 32-33 wk | .0187 b (.012328, .024932) |
| At 34-36 wk | RR, 0.652 |
| At 37-38 wk | RR, 0.492 |
| Rate of infant mortality after birth | |
| At 32 wk | .0125 b (.009, .017) |
| At 33 wk | RR, 0.984 |
| Rate of infant mortality after birth | |
| At 34 wk | RR, 0.704 |
| At 35 wk | RR, 0.488 |
| At 36 wk | RR, 0.344 |
| At 37 wk | RR, 0.312 |
| At 38 wk | RR, 0.224 |
| Risk of infant death with stillbirth or infant death of co-twin | RR, 15 (10-21) |
a Data are baseline value (range in sensitivity analysis), unless otherwise indicated; relative risk compared with referent value at 32 weeks’ gestation;
| Outcome | P value or RR a |
|---|---|
| Monochorionic fetal death | |
| Between 32 + 0 and 33 + 6 wk | .022 b (0, .036) |
| Between 34 + 0 and 35 + 6 wk | RR, 0.427 |
| Between 36 + 0 and 37 + 6 wk 2,5-7,11,12 | RR, 0.387 |
| Rate of other indications that prompt delivery at 32-33 wk | .062 b (.036, .062) |
| Other indications that prompt delivery | |
| At 33-34 wk | RR, 2.15 |
| At 34-35 wk | RR, 2.5 |
| At 35-36 wk | RR, 3.0 |
| At 36-37 wk | RR, 3.76 |
| At 37-38 wk | RR, 4.95 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
