Objective
The purpose of this study was to assess whether early amniotomy reduces the duration of labor or increases the proportion of subjects who are delivered within 24 hours in nulliparous patients who undergo labor induction.
Study Design
We performed a randomized controlled trial that compared early amniotomy to standard management in nulliparous labor inductions. Inclusion criteria were nulliparity, singleton, term gestation, and a need for labor induction. Subjects were assigned randomly to early amniotomy (artificial rupture of membranes, ≤4 cm) or to standard treatment. There were 2 primary outcomes: (1) time from induction initiation to delivery and (2) the proportion of women who delivered within 24 hours.
Results
Early amniotomy shortens the time to delivery by >2 hours (19.0 vs 21.3 hours) and increases the proportion of induced nulliparous women who deliver within 24 hours (68% vs 56%). These improvements in labor outcomes did not come at the expense of increased complications.
Conclusion
Early amniotomy is a safe and efficacious adjunct in nulliparous labor inductions.
Rates of labor induction are rising. Recent data from the National Center for Health Statistics demonstrate an induction rate of >22% in 2006, which was more than double what it was in 1990 and impacted >900,000 births in the United States that year. Although many recent studies have offered evidence for the improvement in methods for labor induction, the induction of labor remains a significant risk factor for cesarean delivery, which highlights the critical need for additional tools to refine induction practice.
Amniotomy, generally thought to be “low-tech,” inexpensive, and safe, has received little research attention, and studies to date have neglected to investigate the efficacy of amniotomy in nulliparous women, despite the fact that more inductions are performed in nulliparous women than their multiparous peers. Based on clinical trials from those in spontaneous labor, early amniotomy timing in labor inductions may shorten the duration of labor. Clinical concern for the rare complications of umbilical cord prolapse and theoretic concerns that rupturing the membranes earlier will lead to a longer overall duration of rupture of membranes with potentially increased rates of chorioamnionitis, neonatal sepsis, and neonatal intensive care unit (NICU) admission have limited the empiric use of the practice of amniotomy in nulliparous labor inductions without level I evidence to support it.
The specific aim of this study was to assess whether early amniotomy, defined as artificial rupture of the membranes, at ≤4-cm dilation, reduces the duration of labor or increases the proportion of subjects delivered within 24 hours in term nulliparous patients who undergo labor induction. We also sought to assess the safety of early amniotomy, as measured by adverse obstetric outcomes and measures of maternal and neonatal infectious morbidities.
Methods
We performed an unblinded, randomized controlled trial at Washington University in St. Louis and the University of Pennsylvania with approval from the Institutional Review Boards at both institutions. Inclusion criteria for this clinical trial were nulliparity, singleton, term gestation (defined as >37 weeks 0 days), and a need for labor induction as determined by the treating physician. Exclusion criteria included HIV infection or cervical dilation of >4 cm at admission examination.
Eligible subjects were approached by trained research nurses and were offered enrollment into the clinical trial. Patients who consented were then randomly assigned to early amniotomy, which was defined as artificial rupture of the membranes at ≤4 cm or to standard management, which was amniotomy at >4 cm dilation. In the early amniotomy group, amniotomy was performed as early as could be done safely. Decisions about the exact timing of rupture (after random assignment) were made by a team that included residents, fellows, and attendings. There were no specific instructions given regarding the timing of amniotomy in the standard treatment group; this decision was left to the treating physicians. The primary method of induction was at the discretion of the treating physician, as were all other intrapartum/postpartum decisions. Random assignment was accomplished centrally. A permuted block randomization procedure was used to formulate assignment lists to assure close to equal numbers of subjects in each treatment group. A uniform block size of 4 was used.
There were 2 primary outcomes. The first was time from initiation of induction, defined as time at delivery of the first induction method to delivery. The second was the proportion of women delivered within 24 hours from the initiation of induction. Although it may seem unusual to have 2 primary endpoints, we believed that both were equally clinically relevant and should be treated as primary outcomes. This decision was made a priori. We also assessed a number of secondary endpoints that included cesarean delivery rates and indications for cesarean delivery, chorioamnionitis (oral temperature >38°C during labor), postpartum fever (oral temperature >38°C on 2 separate occasions >6 hours apart, >24 hours from delivery), wound infection (defined as purulent discharge from the incision), endomyometritis (defined as fundal tenderness and fever that require treatment with antibiotics), NICU admission, and suspected neonatal sepsis. Trained research nurses collected all baseline information, information on the course of labor, and information on maternal and neonatal outcomes.
Statistical analyses were performed with the intent-to-treat principle. Continuous outcomes were compared with the use of the Student t test or Mann-Whitney U dependent on their distributions; dichotomous outcomes were assessed with χ 2 tests or Fisher exact test where appropriate. Time to delivery was not normally distributed and was compared with the use of the Mann-Whitney U test. Relative risks by group and 95% confidence intervals (CIs) were estimated for the percent of women who delivered within 24 hours and each of the secondary outcomes. Our sample size estimate was based on 1 of our primary outcomes: the proportion of women who delivered within 24 hours. We assumed an alpha error of .05, a beta error of .20 (or 80% power), an incidence of delivery within 24 hours of 50% based on published data, a minimum detectable relative risk of 0.75, and a 1:1 allocation ratio. With these assumptions in mind, we estimated that we would need 290 subjects per group. This strategy for sample size calculation gave us tremendous power for our second primary outcome, time to delivery (a continuous measure). Specifically, we estimated a priori that we had 95% power to detect a 2-hour reduction in time to delivery.
Results
Seven hundred forty-nine women were screened for eligibility; 84 women (11.2%) were deemed ineligible by exclusion criteria. Of the 635 eligible nulliparous women, 585 women (92%) consented and were assigned randomly ( Figure ) ; 292 women were assigned to the early amniotomy group, and 293 women were assigned to standard treatment. Those who agreed to participate and those who did not were similar in terms of baseline characteristics (age, gestational age, preexisting medical conditions). The groups were well-balanced with regards to demographics and maternal medical conditions; the mean gestational age at induction was similar between the groups ( Table 1 ). The admission cervical dilation was also similar between the groups. Likewise, the indications for labor induction were similar between the groups. The 2 most common indications for induction were >40-week gestations and gestational hypertension/preeclampsia. The “other” category for indication for induction had a variety of uncommon indications for induction, which included maternal request/social factors (eg, distance from hospital) and oligohydramnios ( Table 2 ).
| Variable | Early amniotomy (n = 292) | Standard therapy (n = 293) | P value |
|---|---|---|---|
| Maternal age, y a | 22.7 ± 5.8 | 23.3 ± 6.2 | .17 |
| African American race, % | 72 | 68 | .30 |
| Diabetes mellitus, % | 3.9 | 3.5 | .81 |
| Chronic hypertension, % | 7.2 | 5.2 | .32 |
| Body mass index, kg/m 2 a | 28 ± 4.2 | 28 ± 3.9 | .90 |
| GBS+, % | 29 | 30 | .66 |
| Pitocin, % | 93 | 93 | .87 |
| Misoprostol, % | 67 | 69 | .70 |
| Cervidil, % | 6.8 | 8.4 | .45 |
| Foley bulb, % | 27 | 30 | .43 |
| More than 1 agent, % | 73 | 73 | .80 |
| Epidural anesthesia, % | 92 | 94 | .88 |
| Admission dilation, cm a | 1.1 ± 1.03 | 1.1 ± 0.97 | .54 |
| Dilation at rupture of membranes, cm | 3.2 | 7.4 | .001 |
| Station at rupture of membranes a | −1 ± 1.2 | −1 ± 1.5 | .50 |
| Cervical examinations, n a | 6.2 ± 3.0 | 5.9 ± 3.4 | .67 |
| Birthweight, g a | 3323 ± 516 | 3311 ± 566 | .78 |
| Gestational age, wk a | 39.7 ± 1.4 | 39.5 ± 1.4 | .16 |
| Variable | Early amniotomy, % | Standard, % | P value |
|---|---|---|---|
| >40 wk | 40 | 39 | .83 |
| Maternal medical indication | 13 | 14 | .72 |
| Gestational hypertension/preeclampsia | 29 | 27 | .63 |
| Intrauterine growth restriction | 6 | 7 | .63 |
| Other | 12 | 13 | .63 |
Methods for induction were similar between the early amniotomy and standard treatment groups. Most women received misoprostol; approximately 30% of the women received a Foley bulb. The induction method categories are not mutually exclusive, because many women received multiple agents. In fact, 73% of women in both groups received >1 agent for induction. As expected, with regards to timing of amniotomy, the early amniotomy group was ruptured earlier than the standard treatment group. Twenty-two women who were assigned randomly to early amniotomy were ruptured after 4 cm of dilation; 13 women who were assigned randomly to standard treatment were ruptured at <4 cm.
The primary results of this study are given in Table 3 . The average time from the start of induction to delivery was shortened by slightly >2 hours in the early amniotomy group (19.0 vs 21.3 hours, respectively; P = .04) compared with those with standard treatment. This difference in length of labor occurred mainly in the first stage of labor, which was defined as time from random assignment to complete cervical dilation. A higher proportion of women in the early amniotomy group were delivered within 24 hours of the initiation of induction (68% vs 56%, respectively; P = .002). Despite these differences in length of labor and delivery within 24 hours, there was no difference in the rate of cesarean deliveries. The rate of chorioamnionitis was increased numerically in the early amniotomy group (11.5% v. 8.5%, respectively; P = .22), although this difference was not statistically significant. Likewise, there were 2 cord prolapses in the early amniotomy group, and none in the standard treatment group.
