Antenatal Corticosteroid Treatment During the Late-Preterm Period and Neonatal Outcomes for Twin Pregnancies

Key Points Question Is there evidence for an association of antenatal corticosteroid treatment during the late-preterm period with an improvement in twins’ neonatal outcomes? Findings In this cohort study, antenatal corticosteroid treatment of 1974 individuals with twin pregnancies and at risk for late preterm birth did not significantly reduce the risk of neonatal respiratory morbidity as well as other neonatal complications. Meaning The evidence is insufficient that antenatal corticosteroid treatment during the late-preterm period in twin pregnancies is associated with a lower risk of newborn morbidity.


Introduction
Over the past 4 decades, the prevalence of twin pregnancies has increased substantially and globally. 1,2Meanwhile, the incidence of preterm births for twin pregnancies has remained notably high. 3,4[7] Antenatal corticosteroids are generally administered to pregnant women at risk for premature delivery between 24 weeks and 0 days and 33 weeks and 6 days of gestation during both singleton and twin pregnancies to accelerate fetal lung maturation and decrease the incidence of neonatal mortality and morbidity. 80][11] As for twins in the late-preterm period, however, the evidence was insufficient to guide clinical practice.Only a few cohort studies, which were characterized by small sample sizes, unmatched comparisons, and great heterogeneity, have been reported.The results showed that the use of antenatal corticosteroid treatment was not consistently associated with an improving neonatal respiratory outcome. 12,13erefore, the correlation between antenatal corticosteroid treatment for twins at risk of latepreterm birth and neonatal outcomes remains an unanswered question.
To better assess the association between antenatal corticosteroid treatment during the latepreterm period and neonatal outcomes of twin pregnancies, we conducted a retrospective cohort study of 2705 twin pregnancies.We used propensity score overlap weighting to make the maternal and neonatal baseline information comparable.Furthermore, to ensure the robustness of the findings, we performed subgroup analysis on potential interactions with the specific variables, including chorionicity, at least 1 infant small for gestational age (SGA), intertwin growth discordance, and infant sex, and we performed sensitivity analysis using propensity score matching and a different administration-to-birth interval and treating twin infants as individuals.

Study Participants
The retrospective cohort involved all individuals whose twin pregnancies were delivered from February 1, 2013, to September 30, 2020, at the Obstetrics and Gynecology Hospital of Fudan University, Shanghai, China, a tertiary hospital.The individuals had been identified from the institution's electronic medical record system, and their electronic records were extracted for the relevant data on maternal characteristics, pregnancy outcomes, neonatal outcomes, and antenatal corticosteroid treatment.Gestational age (GA) was calculated based on the last menstrual period, adjusted by the crown-rump length in the first trimester or the head circumference of the larger fetus with fetal ultrasonographic examination after 14 weeks of gestation in the case of spontaneous conception, as well as based on the timing of in vitro pregnancy fertilization conceived via assisted reproductive technology.Our study conformed to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cohort studies.The study was approved by the Ethics Committee of Obstetrics and Gynecology Hospital of Fudan University, and the requirement for informed consent was waived because the data were deidentified.
The inclusion criteria were as follows: (1) establishment of a prenatal medical record at our hospital and (2) receipt of the routine prenatal care offered by our hospital until delivery.The exclusion criteria were as follows: (1) GA of less than 34 weeks at delivery; (2)   1).

Exposure
The exposure was set as follows: at least 1 dose of antenatal corticosteroids was administered between 34 weeks and 0 days and 36 weeks and 6 days of gestation, in view of a high probability of delivery in the late-preterm period.The protocol for antenatal corticosteroid treatment comprises 4 intramuscular doses of 5 mg of dexamethasone 12 hours apart. 14,15The high probability of delivery was defined as the spontaneous rupture of the membranes, intact membranes and at least 3 cm of dilation or 75% cervical effacement, or a preterm delivery expected for any other indication for which the attending physician decides to induce labor or perform cesarean delivery within 24 hours and 7 days.After birth, maternal characteristics and neonatal outcomes were compared between twins in 2 study groups: those who received at least 1 dose of dexamethasone from 34 weeks and 0 days to 36 weeks and 6 days of gestation and those who did not.

Outcomes
The primary outcome measure was composite neonatal respiratory morbidity, defined as at least a Individuals could fulfill more than 1 exclusion criterion.

Statistical Analysis
Data were analyzed from June 30 to July 13, 2023.The baseline characteristics of the 2 study groups were compared.Continuous variables with normal distributions were presented as mean (SD) values, and categorical data were presented as numbers and percentages.The F test was performed for the comparison of quantitative data, and the χ 2 test or the Fisher exact test was performed for the comparison of categorical data.
For adjusting between-group differences, propensity scores were developed to reflect the probability of each participant receiving antenatal corticosteroid treatment, with the use of multivariable logistic regression with exposure (antenatal corticosteroid treatment vs no antenatal corticosteroid treatment) as the dependent variable and all observed maternal-level and neonatallevel characteristics as the independent variables.The covariates covered baseline maternal sociodemographic and clinical factors, as well as neonatal characteristics, which were expected to be associated with neonatal outcomes.These covariates included year of delivery (2013-2016 and 2017-2020, as the SMFM and the ACOG altered their guidelines regarding antenatal corticosteroid treatment in 2016 to include individuals at risk for late-preterm delivery 9,10 ), advanced maternal age (Ն35 years at delivery), primipara, assisted reproductive technology, gestational diabetes, hypertensive disorders during pregnancy, GA at delivery, preterm birth, mode of delivery (vaginal or cesarean), chorionicity (monochorionicity or dichorionicity), mean birth weight of both twin infants from the same mother, at least 1 infant SGA, intertwin growth discordance, and infant sex (the same or opposite).
Based on propensity scores, 4 adjustment approaches were used to match the baseline of 2 groups, including inverse probability of treatment weighting, standardized mortality ratio weighting, overlap weighting, and propensity score matching (PSM).Inverse probability of treatment weighting, standardized mortality ratio weighting, and overlap weighting, as well as the approach of PSM of 1 to 1, 2, 3, and 4 matching patients with the nearest neighbor matching algorithm (a caliper width of 0.1), were performed to examine the association between antenatal corticosteroid treatment and neonatal outcomes.Standardized mean differences were used to select the optimal adjustment method.As the absolute standardized mean difference of all covariates was smaller than 10% to support the assumption of balance between study groups, 16 overlap weighting was chosen for subsequent research (eFigure 1 in Supplement 1).The outcomes were investigated using univariate logistic regression.Odds ratios (ORs) with 95% CIs were used to show the effect size of a comparison.
We conducted 6 subgroup analyses of the primary outcome to examine the heterogeneity in clinically relevant subgroups: GA at delivery (34 weeks and 0 days to 34 weeks and 6 days, 35 weeks and 0 to 35 weeks and 6 days, and 36 weeks and 0 days to 36 weeks and 6 days), year of delivery, chorionicity, at least 1 infant with SGA, intertwin growth discordance, and infant sex.The robustness of the results was examined via multiple sensitivity analyses.First, we performed 1:1 PSM as an alternative to overlap weighting because the absolute standardized mean difference of almost all covariates was smaller than 10% (eFigure 1 in Supplement 1).Second, we regarded twin neonates as individuals, adding birth order, infant sex, and birth weight to the covariates.In addition, we further divided the antenatal corticosteroid treatment group into 3 groups based on the time from the administration of the first dose of dexamethasone to birth-less than 24 hours, 24 hours to 7 days, and more than 7 days-and used multivariate logistic regression analyses to assess whether the administration-to-birth interval was a factor in the association of antenatal corticosteroid treatment with neonatal outcomes.
A 2-tailed P < .05 was considered statistically significant.All statistical analyses were conducted with R statistical software, version 4.2.2 (R Project for Statistical Computing).

Results
A total of 2705 individuals with twin pregnancies were eligible for the current retrospective analysis, of whom 731 met the exclusion criteria.2).Subgroup analysis showed no significant interaction for the neonatal primary outcomes between antenatal corticosteroid treatment and other factors, such as GA at delivery, year of delivery, chorionicity, at least 1 infant with SGA, intertwin growth discordance, and infant sex (Figure 2).
The sensitivity analyses showed a similar result in terms of the association of antenatal corticosteroid treatment with neonatal outcomes after 1:1 PSM (eTables 2 and 3 in Supplement 1) and after treating twin neonates as separate individuals (eTables 4 and 5 in Supplement 1).From the multivariable logistic regression analysis to investigate the association of the administration-to-birth interval with the association of antenatal corticosteroid treatment with neonatal outcomes, no differences were observed in the risk of neonatal primary outcomes at different intervals.As to the secondary neonatal outcomes, the risk of neonatal hypoglycemia was significantly higher in the  a Definitions of outcome measures listed in eTable 1 in Supplement 1.
b The primary outcome measure was a composite neonatal respiratory morbidity outcome defined as at least 1 of the following occurrences in at least 1 neonate of the twins: respiratory distress syndrome, mechanical ventilation, surfactant administration, transferred with respiratory complications, or neonatal death.

Discussion
In this large, single-center retrospective cohort study, we did not find evidence that antenatal corticosteroid treatment of individuals with twin pregnancy at risk for late-preterm birth could reduce the risk of neonatal respiratory morbidity.The finding was confirmed by the subgroup interaction tests, including GA at delivery, year of delivery, chorionicity, at least 1 infant with SGA, intertwin growth discordance, and infant sex, as well as sensitivity analysis using propensity score matching and a different administration-to-birth interval and treating twin infants as individuals.
However, the risk of neonatal hypoglycemia increased significantly when antenatal corticosteroid treatment was administered 1 to 7 days prior to birth, as was previously reported. 17ere has been considerable evidence supporting antenatal corticosteroid treatment during the late-preterm period for singleton pregnancies 11,18,19 but limited, unsupportive evidence for antenatal corticosteroid treatment for twin pregnancies. 12,13This difference could be explained by the differences in the pharmacokinetics of corticosteroids in twins vs singletons, the off-target phenomenon of antenatal corticosteroids (the absence of delivery Յ7 days after administration of antenatal corticosteroids), or the confounding factors associated with antenatal corticosteroid choice.
To minimize the confounding factors in antenatal corticosteroid treatment of individuals with twin pregnancy, we implemented the following strategies.We took advantage of a cohort of 2705 individuals with twin pregnancies, which, to our knowledge, is larger than that of previously reported studies. 12,13Instead of the conventional multivariable logistic regression previous studies used, we used the propensity score overlap weighting method to effectively eliminate the potential confounding factors, achieving the effect of randomized clinical trials to the greatest extent possible in retrospective studies. 20For outcome measures, based on earlier studies, such as the ALPS trial and the clinical realities of our center, we developed outcome indicators that could comprehensively reflect the association between antenatal corticosteroid treatment and neonatal outcomes among twins. 11As there are no randomized clinical trials on the effect of antenatal corticosteroid treatment in twin pregnancy to our knowledge, we believe that carefully designed observational studies can be of great help in treatment decision-making.Next, the results did not change significantly in the multiple subgroup and sensitivity analyses, showing that the results were robust enough to be convincing and reliable.
2][23][24] Unlike the 2 previously reported investigations, 12,13 our study aimed to explore the potential association of the administration-to-birth interval with outcomes; we found that antenatal corticosteroid treatment, regardless of the length of the interval, was not associated with improved neonatal respiratory outcomes, which suggested that our findings were not associated with the off-target phenomenon.
The lack of benefit associated with antenatal corticosteroid treatment in twin pregnancies may be attributed to the different pharmacokinetics of corticosteroids between twins and singletons.
Corticosteroids are thought to have a shorter half-life and a higher clearance rate among twins, and the physiological changes in the maternal blood volume during twin pregnancy may affect drug distribution, all of which may lead, for twins, to lower effective therapeutic doses of corticosteroids that are intended to promote fetal lung maturation. 25,26Although the beneficial association of corticosteroids with respiratory outcomes for early-preterm singleton infants is widely recognized, controversy still exists over their association with outcomes for early-preterm twin infants, with large-sample studies demonstrating no improvement in respiratory outcomes. 27,28Hence, the corticosteroid dosage currently in use is probably not optimal enough to promote fetal lung maturation in twin pregnancies.
Our study suggested that antenatal corticosteroid treatment for twin pregnancies at risk of latepreterm birth could not decrease neonatal respiratory morbidity but could increase the risk of neonatal hypoglycemia when administered 1 to 7 days prior to birth.Thus, antenatal corticosteroids should not be administered to those who do not have a high likelihood of preterm birth with twin pregnancies until more robust evidence is available.The current shortcomings in the use of antenatal corticosteroid treatment for twins should be properly addressed, such as the lack of data on optimal dosage, estimating preterm birth, and the long-term effect on multiple systems.Further randomized clinical trials are required to help determine optimal antenatal corticosteroid treatment recommendations for this special risk group of women.

Limitations
This study has several limitations.First, this cohort study performed a retrospective observational analysis.Although we used propensity score overlap weighting to balance the differences between groups, unmeasured and residual confounding could still exist.Second, although one of the primary outcome indicators was patient transferred with respiratory complications, the information on the prognosis of these newborns was inevitably lost, which increased the potential for bias.Third, due to the challenges in estimating the timing of delivery, only approximately 30% of the antenatal corticosteroid treatment group completed the full course of corticosteroids, which may have led to bias in the outcomes.Fourth, the findings need to be interpreted with caution because some neonatal outcomes, such as death in the late-preterm period, had a low incidence.Fifth, the cohort study spanned approximately 8 years, during which the changes in clinical practice, medical standards, and policy guidelines could have exerted a potential effect on the generalizability of the conclusion.

Conclusion
In this cohort study, evidence was insufficient that antenatal corticosteroid treatment for a twin pregnancy during the late-preterm period could be associated with a lower risk of newborn morbidity.However, antenatal corticosteroid treatment 1 to 7 days before birth was associated with a higher risk of neonatal hypoglycemia.More well-designed prospective studies are required.

Table 1 .
). Baseline Characteristics of the Study CohortsAfter overlap weighting, the univariate logistic regression analysis showed no significant differences between the antenatal corticosteroid treatment group and the no-antenatal corticosteroid treatment group in the risk of neonatal primary outcome (29 of 303 [9.6%] vs 41 of 1671 [2.5%]; weighted OR, 1.27 [95% CI, 0.60-2.76]).Similar results were shown for the risk of being transferred with respiratory complications, respiratory distress syndrome, and surfactant administration.The risks of other neonatal complications, such as CPAP use, CPAP use for more than a Other indications: breech presentation, shoulder presentation, placenta accrete, placenta previa, intrahepatic cholestasis of pregnancy, monochorionic twins, or others.JAMA Network Open | Obstetrics and GynecologyAntenatal Corticosteroid Treatment and Neonatal Outcomes for Twin Pregnancies JAMA Network Open.2023;6(11):e2343781. doi:10.1001/jamanetworkopen.2023.43781(Reprinted) November 17, 2023 5/11 Downloaded from jamanetwork.comby guest on 11/25/2023 corticosteroid treatment group than in the no-antenatal corticosteroid treatment group at 0.49-0.84])were significantly lower in the antenatal corticosteroid treatment group than in the no-antenatal corticosteroid treatment group at the interval of less than 24 hours and more than 7 days, respectively (eFigure 2 and eTables 6, 7, and 8 in Supplement 1).

Table 2 .
Association of ACS Administration With Neonatal Outcomes Abbreviations: ACS, antenatal corticosteroid; CPAP, continuous positive airway pressure; NA, not applicable; NICU, neonatal intensive care unit; OR, odds ratio.
1. eTable 1. Definitions of Variables and Outcome Measures eFigure 1. Methods Compared eTable 2. Baseline Characteristics of the Study Cohorts After 1:1 Propensity Score Matching eTable 3. Association of ACS Administration With Neonatal Outcomes After 1:1 Propensity Score Matching eTable 4. Baseline Characteristics of the Study Cohort (Neonates) eTable 5. Association of ACS Administration With Neonatal Outcomes After Overlap Weighting (Neonates) eFigure 2. Description of the Study Design for the Analysis of ACS Administration-to-Delivery Intervals eTable 6. Baseline Characteristics of the Study Cohorts at Different Intervals of ACS Administration-to-Birth eTable 7. Association of ACS Administration With Primary Neonatal Outcomes at Different Intervals of ACS Administration-to-Birth After Adjustment eTable 8. Association of ACS Administration With Other Secondary Neonatal Outcomes at Different Intervals of ACS Administration-to-Birth After Adjustment