The results of the recent U.S. multicenter randomized trial of elective induction at 39 weeks of gestation compared with expectant management (ARRIVE [A Randomized Trial of Induction Versus Expectant Management]) revealed that elective induction of labor at 39 weeks in low-risk, nulliparous participants did not significantly lower the frequency of a composite perinatal morbidity or death, but lowered the frequencies of cesarean delivery, hypertensive disorders of pregnancy, and neonatal respiratory morbidity.1,2 As a result, the American College of Obstetricians and Gynecologists and the Society for Maternal-Fetal Medicine released recommendations suggesting that it is reasonable to offer induction of labor to low-risk nulliparous patients at 39 weeks of gestation.3,4 Therefore, the rates of induction of labor in the United States are expected to rise significantly above current rates of one in four births.5 As these guidelines are implemented, it is useful to health care professionals and low-risk pregnant patients to have information regarding the natural course and outcomes of expectant management to optimize shared decision-making.6 However, there is a paucity of contemporary, reliable and prospectively collected information regarding the ongoing risks of expectant management at 39 weeks of gestation among low-risk patients. Such data are essential to provide further insight into the study findings and to assist in the appropriate implementation of the new recommendations. Therefore, we used data from the ARRIVE trial to estimate the risks of cesarean delivery, the perinatal composite outcome, and other outcomes of interest by completed week of gestation after 39 weeks in low-risk nulliparous patients who participated in the trial and underwent expectant management. We also describe the frequency of medically indicated induction of labor.
This was a secondary analysis of ARRIVE, the multicenter open-label randomized trial of elective induction of labor at 39 0/7–39 4/7 weeks of gestation compared with expectant management until at least 40 5/7 weeks but no later than 42 2/7 weeks in low-risk nulliparous patients. Details of the trial have been previously reported.1 Briefly, nulliparous patients were screened at 34–38 completed weeks of gestation at 41 U.S. hospitals participating in the Eunice Kennedy Shriver National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network and randomized at 38 0/7–38 6/7 weeks. Inclusion criteria included live cephalic singleton pregnancy, absence of a contraindication to vaginal delivery, no plan for cesarean delivery or labor induction at the time of randomization, and reliably determined gestational age consistent with American College of Obstetricians and Gynecologists guidelines. A reliable gestational age was defined by a sure last menstrual period consistent with an ultrasonogram before 21 0/7 weeks of gestation or by an ultrasonogram before 14 0/7 weeks if last menstrual period was uncertain. Exclusion criteria included the presence of any maternal or fetal indication for delivery before 40 5/7 weeks of gestation (eg, hypertension, diabetes mellitus, fetal growth restriction, placental abruption). Trained and certified research staff abstracted information from medical records, including demographic information, medical, obstetric, and social history, and outcome data. Institutional review board approval was available at each of the participating sites and written informed consent was obtained from participants before randomization.
The current analysis included only participants who were randomized to and underwent expectant management and attained 39 0/7 weeks of gestation. Participants without delivery outcomes or who delivered newborns with major fetal anomalies were excluded. For this analysis, delivery gestational age was categorized by completed week (39 weeks: 39 0/7–39 6/7 weeks, 40 weeks: 40 0/7–40 6/7 weeks and 41–42 weeks: 41 0/7–42 2/7 weeks; none delivered after 42 2/7 weeks).
The primary maternal outcome for this analysis was cesarean delivery. Other maternal outcomes examined were hypertensive disorders of pregnancy (preeclampsia or gestational hypertension), peripartum infection (chorioamnionitis and endometritis), placental abruption, shoulder dystocia and an adverse composite outcome comprising blood transfusion, B-lynch suture, uterine artery ligation, hypogastric artery ligation, vascular embolization, hysterectomy, or intensive care unit admission.
The primary perinatal outcome for this analysis was the same perinatal composite as was used for the primary outcome of the ARRIVE trial: perinatal death, the need for respiratory support within 72 hours after birth, 5-minute Apgar score 3 or less, hypoxic ischemic encephalopathy, seizure, infection (confirmed sepsis or pneumonia), meconium aspiration syndrome, birth trauma (bone fracture, neurologic injury, or retinal hemorrhage), intracranial or subgaleal hemorrhage, or hypotension requiring vasopressor support. Respiratory support was defined as use of continuous positive airway pressure, high-flow (more than 1 L/min) nasal cannula, mechanical ventilation, or cardiovascular resuscitation within the first 72 hours of life. Hypoxic ischemic encephalopathy was defined according to Shankaran criteria.7 Secondary neonatal outcomes included neonatal intermediate or intensive care unit admission, as well as a composite of neonatal death or stillbirth. We also estimated the frequency of medically indicated delivery and specific indications.
Categorical variables were compared by completed week of gestation at delivery using Cochran-Armitage trend test, or Monte-Carlo estimate for exact trend test for rare outcomes, and continuous variables using a Jonckheere Terpstra test. For multivariable analysis, we applied log-binomial regression and modified Poisson regression and computed the adjusted relative risk (aRR) and 95% CI for each week of gestation (40 0/7–40 6/7 and 41 0/7–42 2/7 weeks vs 39 0/7–39 6/7 weeks). Models for the maternal outcomes were adjusted for maternal age, smoking, and body mass index (BMI, calculated as weight in kilograms divided by height in meters squared) at first clinic visit. Models for the perinatal and neonatal outcomes were adjusted for maternal age, smoking, BMI at first clinic visit, and fetal sex. We further examined whether the association between completed gestational age at delivery and outcomes were modified by cervical status (modified Bishop score less than 5 compared with 5 or higher) at randomization, BMI (less than 30, 30–39, and 40 or higher), or participant-reported race–ethnicity (interactions). Alternative analyses evaluated outcomes among those who attained 39 0/7 weeks of gestation (those who delivered after 39 6/7 weeks were compared with those who delivered at 39 0/7–39 6/7 weeks) and among those who attained 40 0/7 weeks (those who delivered after 40 6/7 weeks were compared with those who delivered at 40 0/7–40 6/7 weeks). We chose not to adjust for race–ethnicity and insurance type in primary models as these are social determinants of health and did evaluate race–ethnicity in supplementary analyses. Statistical analyses were performed with SAS 9.4. A two-sided P<.05 was considered to indicate statistical significance for descriptive and unadjusted analyses. For multivariable analysis, P<.0125 was considered to indicate statistical significance for the coprimary outcomes and for the secondary outcomes, the level of significance was adjusted post hoc for multiple comparisons with the false discovery rate method.8 No imputation for missing data was performed.
Of 6,106 participants, 3,037 of the 3,044 randomized to expectant management had outcome data. Of these, a total of 2,502 (82.4%) participants met eligibility criteria for this analysis (Fig. 1). Excluded were 274 (9.0%) participants owing to spontaneous labor, 115 (3.8%) with indicated delivery before 39 0/7 weeks, 139 (4.6%) electively delivered rather than expectantly managed after 39 0/7 weeks, and seven (0.2%) owing to major anomalies detected after delivery.
Of the 2,502 participants analyzed, 964 (38.5%) delivered at 39 0/7–39 6/7 weeks of gestation, 1,111 (44.4%) at 40 0/7–40 6/7 weeks, and 427 (17.1%) at 41 0/7–42 2/7 weeks; only 8 of the 427 delivered at 42 0/7–42 2/7 weeks. Selected demographic characteristics at randomization differed by completed week of gestation at delivery (Table 1). The frequencies of modified Bishop score less than 5 (60.6–81.7%) and smoking (6.2–11.0%) and the medians of maternal age (23–24 years) and BMI (29.9–31.6) increased as delivery gestational age increased from 39 to 41–42 completed weeks (Table 1). Fewer participants who delivered at 41 0/7–42 2/7 weeks of gestation had private insurance compared with those who delivered earlier.
The type of labor and reasons for delivery differed by completed week of gestation (Table 2). Spontaneous labor occurred in 1,552 (62.0%) participants overall, decreasing from 76.2% among those who delivered at 39 0/7–39 6/7 weeks of gestation to 19.7% among those who delivered at 41 0/7–42 2/7 weeks. The proportion of women who had a cesarean delivery among those with spontaneous labor increased as gestational age advanced. The corresponding frequency of all medically indicated deliveries (including for the indication of postterm) was 38.0% (n=950); 96.6% (n=918) were by induction and 3.4% (n=32) were by cesarean delivery without labor. The proportion of medically indicated deliveries increased from 23.8% for delivery at 39 0/7–39 6/7 weeks of gestation to 80.3% at 41 0/7–42 2/7 weeks. Among all reasons for medically indicated deliveries, postterm was the most common; this indication comprised 36.2% of medical indications at 40 0/7–40 6/7 weeks of gestation (specifically at 40 5/7 weeks and 40 6/7 weeks allowed by the study protocol) and 88.3% of medical indications at 41 0/7–42 2/7 weeks (Table 2). Other than postterm induction (at 40 completed weeks of gestation), hypertensive disorders of pregnancy and premature membrane rupture were the most common reasons for delivery at 39 or 40 completed weeks.
Among maternal and perinatal outcomes, the frequency of cesarean delivery (17.3%, 22.0%, 37.5%; P<.001) and the adverse perinatal composite outcome (5.1%, 5.9%, 8.2%; P=.03) increased significantly with 39, 40, and 41–42 completed weeks of gestation, respectively, whereas the frequency of hypertensive disorders of pregnancy decreased significantly (16.4%, 12.1%, 10.8%, P=.001) (Table 3). No significant changes as a function of gestational age were seen in the frequency of the maternal adverse composite outcome, placental abruption, peripartum infection, or neonatal intermediate or intensive care unit admission.
The adjusted results when comparing outcomes of deliveries at 40 0/7–40 6/7 weeks of gestation and 41 0/7–42 2/7 weeks with 39 0/7–39 6/7 weeks as the referent were mostly consistent with the unadjusted results (Table 4). There were significant increases in the risk of cesarean delivery at 41 0/7–42 2/7 weeks of gestation (aRR 1.93, 95% CI 1.61–2.32), but not the perinatal composite (after adjustments), and decreases in the risk of hypertensive disorders of pregnancy at 40 0/7–40 6/7 weeks (aRR 0.71, 95% CI 0.58–0.88) and 41 0/7–42 2/7 weeks (aRR 0.61, 95% CI 0.45–0.82). Other maternal and perinatal outcomes were not significantly associated with gestational age at delivery. The numbers of participants experiencing placental abruption were too few for multivariable analysis.
Inclusion of race–ethnicity in the multivariable analysis did not change the results. No interactions between gestational age and modified Bishop score, BMI, or race–ethnicity were found to be significant for any of the study outcomes (all P>.05).
In the alternative analyses, among participants who reached 39 0/7 weeks of gestation undelivered, the risk of cesarean delivery was higher among those who delivered after 39 6/7 weeks compared with those who delivered at 39 0/7–39 6/7 weeks (aRR 1.45, 95% CI 1.24–1.71), whereas the risk of hypertensive disorders of pregnancy was lower (aRR 0.68, 95% CI 0.56–0.83) (Table 5). Among participants who reached 40 0/7 weeks of gestation undelivered, the risk of cesarean delivery was higher among those who delivered after 40 6/7 weeks compared with those who delivered at 40 0/7–40 6/7 weeks (aRR 1.56, 95% CI 1.33–1.84).
Among nulliparous participants who were expectantly managed beyond 39 0/7 weeks of gestation, we found an increasing risk of both an adverse composite perinatal outcome and cesarean delivery with increasing gestational age. However, after adjustments in multivariable analysis, the risk of cesarean delivery was 25% higher at 40 0/7–40 6/7 weeks of gestation and 93% higher 41 0/7–42 2/7 weeks compared with 39 0/7–39 6/7 weeks. The perinatal composite outcome was 56% higher at 41 0/7–42 2/7 weeks of gestation but did not meet criteria for statistical significance. In contrast, the prevalence of hypertensive disorders of pregnancy decreased with advancing gestational age. Although two thirds of participants who undergo expectant management will labor spontaneously, the frequency of medically indicated delivery increased with advancing gestational age. Postterm induction, hypertensive disorders of pregnancy, and premature membrane rupture were the most common reasons for medically indicated delivery.
Since the publication of the ARRIVE trial, a meta-analysis of cohort studies of elective induction of labor at 39 0/7–39 6/7 weeks of gestation as compared with expectant management has been reported.9 The results of this meta-analysis revealed that elective induction of labor was associated with a reduction in cesarean delivery, peripartum infection, and adverse perinatal outcomes. These findings lend additional support to the recommendations by professional organizations to offer elective induction of labor to low-risk nulliparous patients.3,4 Our study focusing on participants who underwent expectant management addresses some questions that remained unanswered regarding perinatal outcomes on patients expectantly managed beyond 39 0/7 weeks of gestation.6,10
Limited data were reported after a U.K. randomized trial of induction of labor compared with expectant management involving 619 primigravid participants who were 35 years of age or older, which preceded the ARRIVE trial. Findings from the trial suggested that 46% of participants in the expectant management group had spontaneous labor, 49% had induction of labor, and 2% had cesarean delivery without labor.11 Furthermore, the main reasons for induction of labor were similar.
Our findings are also supported by data from cohort studies that have simulated expectant management6,12–14—most deliveries occur at 39–40 completed weeks of gestation, the most common reasons for inductions are similar, and cesarean delivery increases with advancing gestational age. Although the cesarean delivery rates in nulliparous patients were higher at corresponding gestational ages in other studies compared with the rates in the ARRIVE trial, this likely reflects the inclusion only of low-risk participants in our cohort.11,12 The decrease in hypertensive disorders of pregnancy (in contrast to increase in cesarean delivery) with advancing gestational age after 39 0/7–39 6/7 weeks is surprising and deserves further investigation. For reference, the overall rate of hypertensive disorders of pregnancy in the primary ARRIVE article was 9.1% in the labor induction group and 14.1% in the expectant management group, compared with an overall 13.5% in the expectant management group in this article and the 16.4% at 39 0/7–39 6/7 weeks of gestation, 12.1% at 40 0/7–40 6/7 weeks and 10.8% at 41 0/7–42 2/7 weeks in the expectant management group. The difference is due to the slight variation in inclusion–exclusion criteria for this analysis. A query of the PubMed database using key words preeclampsia, hypertensive disorders and timing of delivery or gestational age of delivery, did not yield any publication reporting such a trend. We do not believe that the incidence of preeclampsia decreases with advancing gestational age. It may reflect an increase in the prevalence of competing indications for delivery such as postterm inductions as gestational age advances (Table 2). Other indications for delivery were more likely to be prevalent than preeclampsia among deliveries with advancing gestational age.
We acknowledge a number of study limitations. Some of the specified outcomes such as stillbirth or neonatal death were too infrequent to conduct any meaningful analyses. In the ARRIVE trial, only nulliparous participants are included, limiting generalizability. We did not account for multiple comparisons or center differences; some of the findings could be due to chance or center practices, such as the association of insurance type with gestational age at delivery. However, no center effect in relation to outcomes was noted in the primary trial.1 There are several strengths of the study. These data are relatively novel as there are limited, if any, such data from large randomized trials. The analysis uses prospectively collected data by certified research staff, enhancing validity. The sample size is large enough to evaluate the key outcomes.
Overall, these data provide important insights into the anticipated course of expectant management of low-risk nulliparous patients who would typically be candidates for elective induction of labor. The information will be useful for counseling patients regarding ongoing risks of expectant management, and optimizing shared decision making. About two thirds of expectantly managed participants on average labored spontaneously after 39 weeks of gestation, but the frequency decreased with advancing gestational age. Conversely, the frequencies of medically indicated induction of labor and cesarean delivery increased with gestational age.
1. Grobman WA, Rice MM, Reddy UM, Tita ATN, Silver RM, Mallett G, et al. Labor induction versus expectant management in low-risk nulliparous women. New Engl J Med 2018;379:513–23. doi: 10.1056/NEJMoa1800566
2. Greene MF. Choices in managing full-term pregnancy. New Engl J Med 2018;379:580–1. doi: 10.1056/NEJMe1807747
3. American College of Obstetricians and Gynecologists. ACOG practice advisory: clinical guidance for integration of the findings of the ARRIVE trial: labor induction versus expectant management in low-risk nulliparous women. Accessed August 10, 2020. https://www.acog.org/clinical/clinical-guidance/practice-advisory/articles/2018/08/clinical-guidance-for-integration-of-the-findings-of-the-arrive-trial
4. Society for Maternal-Fetal Medicine (SMFM) Publications Committee. SMFM statement on elective induction of labor in low-risk nulliparous women at term: the ARRIVE trial. Am J Obstet Gynecol 2019;221:B2–4. doi: 10.1016/j.ajog.2018.08.009
5. Little SE. Elective induction of labor: what is the impact? Obstet Gynecol Clin North Am 2017;44:601–14. doi: 10.1016/j.ogc.2017.08.005
6. Carmichael SL, Snowden JM. The ARRIVE trial: interpretation from an epidemiologic perspective. J Midwifery Women's Health 2019;64:657–63. doi: 10.1111/jmwh.12996.
7. Shankaran S, Laptook AR, Ehrenkranz RA. Whole-body hypothermia for neonates with hypoxic-ischemic encephalopathy. N Engl J Med 2005;353:1574–84. doi: 10.1056/NEJMcps050929
8. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc [B] 1995;57:289–300.
9. Grobman WA, Caughey AB. Elective induction of labor at 39 weeks compared with expectant management: a meta-analysis of cohort studies. Am J Obstet Gynecol 2019;221:304–10. doi: 10.1016/j.ajog.2019.02.046
10. Marrs C, LaRose M, Caughey A, Saade G. Elective induction at 39 weeks of gestation and the implications of a large, multicenter, randomized controlled trial. Obstet Gynecol 2019;133:445–50. doi: 10.1097/AOG.0000000000003137
11. Walker KF, Bugg GJ, Macpherson M. Randomized trial of labor induction in women 35 years of age or older. N Engl J Med 2016;374:813–22. doi: 10.1056/NEJMoa1509117
12. Darney BG, Snowden JM, Cheng YW, Jacob L, Nicholson JM, Kaimal A, et al. Elective induction of labor at term compared with expectant management: maternal and neonatal outcomes. Obstet Gynecol 2013;122:761–9. doi: 10.1097/AOG.0b013e3182a6a4d0
13. Stock SJ, Ferguson E, Duffy A, Ford I, Chalmers J, Norman JE. Outcomes of elective induction of labour compared with expectant management: population based study. BMJ 2012;344:e2838. doi: 10.1136/bmj.e2838
14. Glantz JC. Term labor induction compared with expectant management. Obstet Gynecol 2010;115:70–6. doi: 10.1097/AOG.0b013e3181c4ef96