Materials and Methods

A prospective randomized multicenter trial was conducted by 11 Obstetrix practices and 13 non-Obstetrix practices in 27 hospitals, with 17 of these practices in 18 hospitals enrolling at least 1 subject. The study was approved by institutional review boards at each participating hospital. The study was registered on clinicaltrials.gov ( NCT00201656 ). Patients were eligible for randomization if they were between 22 weeks 0 days and 32 weeks 6 days of pregnancy and if they had a previously placed McDonald or Shirodkar cerclage defined as any cerclage done at least 1 week prior and at ≤23 weeks and 6 days, including those done for history of cervical incompetence or asymptomatic cervical shortening and/or asymptomatic cervical dilation <3 cm. Singletons and twins were included. Patients were not eligible for inclusion if they were in active labor and/or had >8 contractions per hour, had evidence of chorioamnionitis (temperature >38°C plus fetal tachycardia or uterine tenderness), had nonreassuring fetal status by fetal heart rate (FHR) monitoring or biophysical profile, had placenta previa, had evidence of fetal lung maturity or infection by amniocentesis, had major fetal anomaly on ultrasound, or if rupture of membranes had occurred >48 hours prior to randomization. Also ineligible were those whose rupture of membranes followed amniocentesis in the past week or who had an abdominal cerclage placed.

All patients underwent speculum examination to confirm rupture of membranes with presence of at least 2 of the following: pooling, ferning, and a positive Nitrazine test. After confirming PPROM, patients were evaluated to rule out active labor, fetal distress, and chorioamnionitis. Ultrasound was performed to determine gestational age, normal fetal anatomy, and amniotic fluid index. Amniocentesis was performed at the discretion of the managing clinician. Fluid collected by amniocentesis was sent for Gram stain, culture, glucose, and white blood cell count. Management based on the results of amniocentesis when performed was at the discretion of the clinician.

Patients were consented and randomized to immediate removal of cerclage vs retention of cerclage and randomization and removal of cerclage, when performed, was done as soon as possible after randomization. A computer-generated random-number sequence was used to generate randomization assignments that were stratified by institution in blocks of 6, maintained at a central location, and available 7 days a week, 24 hours a day. All patients were then managed expectantly and only delivered for any of the following: labor; chorioamnionitis; documented fetal lung maturity; nonreassuring fetal status; or until the patient reached 34 weeks at which time the cerclage was removed in the group with cerclage remaining. All randomized patients were treated with antibiotics. The suggested regime was ampicillin (2 g intravenous [IV] every 6 hours for 8 doses) and erythromycin (250 mg IV every 6 hours for 8 doses) for 48 hours followed by amoxicillin (250 mg oral dose every 8 hours) and erythromycin (333 mg oral dose every 8 hours) for 5 days of oral therapy. Patients allergic to ampicillin received erythromycin as above, and ampicillin was substituted by clindamycin (900 mg IV every 8 hours for 6 doses) for 48 hours followed by clindamycin (450 mg oral dose every 8 hours) for 5 days of oral therapy. Patients allergic to erythromycin received ampicllin/sulbactam (1.5 g IV every 6 hours for 8 doses) for 48 hours followed by amoxicillin/clavulanate (500 mg oral dose every 8 hours) for 5 days. Patients ≥24 weeks received corticosteroids, 1 complete course only, with betamethasone (12 mg intramuscularly, 2 doses 24 hours apart) to enhance pulmonary maturity (dexamethasone 6 mg IV or intramuscularly, 4 doses, 12 hours apart was used only in hospitals where betamethasone was not available). All patients had continuous FHR monitoring for 12-24 hours (if at least 24 weeks) followed by intermittent FHR monitoring on a daily basis. No tocolytics were permitted to be used in either group. Prior tocolytic treatment was not an exclusion criterion; however, tocolysis was discontinued prior to randomization. Magnesium sulfate was given for neuroprotection as appropriate. Expectant management was followed until the onset of labor, evidence of fetal distress, chorioamnionitis, at which time patients were delivered after removal of cerclage if in that group. Management once 34 weeks was reached was at the discretion of the managing clinician.

The primary hypothesis for this study was that, in the presence of a cervical cerclage and subsequent PPROM, leaving the cerclage in place as compared to removing it would result in prolongation of pregnancy without increasing the risk of infection. Therefore, the primary outcome for this study was prolongation of pregnancy by at least 1 week and second major outcome was the rate of chorioamnionitis between the 2 groups. Secondary outcomes included total days of pregnancy prolongation after randomization; gestational age at delivery; postpartum endometritis; maternal sepsis; and composite neonatal outcome, which included any one of the following (for twins, either infant): fetal or neonatal death, respiratory distress syndrome, documented neonatal sepsis within 72 hours of delivery, grade-3 or -4 intraventricular hemorrhage, and/or stage-2 or -3 necrotizing enterocolitis (NEC). Additionally evaluated outcomes included duration of neonatal intensive care unit stay and birthweight.

The sample size for the study was determined using a power of 80% at a P value of .05 and using a 1:1 case-to-control ratio. We calculated that a sample of 142 patients (71 per group) would yield 80% power to detect a difference in the rate of chorioamnionitis from 22% in the cerclage removal group to 44% in the cerclage retention group and >80% power to detect a difference in prolongation of pregnancy >1 week from 50% in the cerclage removal group to 75% in the cerclage retention groups. P values for maternal baseline characteristics and outcomes were calculated using the Wilcoxon rank sum test for continuous variables and using Fisher exact test for categorical variables. Differences between the treatment groups in neonatal outcomes were determined from a repeated measures model where each baby was considered the repeated measure (generalized estimating equations for dichotomous variables and a mixed linear model for continuous variables). The models also controlled for twin gestation. For NEC and gestational age at delivery, the repeated measures model would not converge. Thus, for NEC the P value was calculated using a Fisher exact test (unadjusted for twin gestation due to the small numbers of babies with NEC). For gestational age at delivery, the P value was determined using a linear regression controlling for twin gestation. Analysis followed the intent-to-treat principle, that is, patients were analyzed according to their assigned group regardless of any subsequent management after randomization.

A data safety monitoring committee (DSMC) was established, consisting of 2 individuals whose institutions did not participate in the study. The DSMC reviewed maternal and neonatal infection rates and other serious adverse outcomes. The DSMC reviewed data after each 25 cases or sooner if any maternal death, serious maternal septic complication, or septic neonatal death occurred. If a statistically significant increase in infection or other trend toward major complications developed in either group, the study was to be terminated after consultation with the investigators. A futility analysis was conducted using the conditional power approach. The probability of finding a statistically significant result at the end of the study was calculated based on the current trend in the data, the null and alternative hypotheses.

Results

The study was conducted from June 2003 through November 2013. After the second interim analysis was performed the DSMC recommended a futility analysis that was performed by an independent statistician. Based on the original sample size of 142, the conditional power of achieving a difference in latency between the groups was determined to be 17% in favor of showing a significant difference in the wrong direction under the current trend, 0.1% under the null hypothesis, and 22.8% under the alternative hypothesis. There was a trend towards increased chorioamnionitis in the retention group (see data below). It was therefore determined by the DSMC, and subsequently agreed upon by the authors, that the original null hypothesis would not have any reasonable likelihood of being rejected even if the originally planned target enrollment of 142 patients was achieved. Therefore it was decided to terminate the study at this point.

At the time of study termination, a total of 58 women were randomized, 32 to removal and 26 to retention of cerclage and the following analyses relate to these 56 patients with outcome data after exclusion of 2 subjects, one of whom withdrew and the other who was lost to follow up. All 56 underwent the assigned management (removal or retention of cerclage) and there was no crossover. The enrollment data are shown in the Figure . Baseline characteristics of each group were generally comparable except for more twins in the cerclage retention group and are shown in Table 1 . Nine patients underwent amniocentesis, 6 in the removal group and 3 in the retention group. None of these patients were delivered on the basis of results of the amniocentesis. As shown in Table 2 , for the primary hypothesis, latency was not prolonged in the retention group. The incidence of being undelivered at 1 week was 56% for the removal group and 45% for the retention group ( P = .58). The mean latency for the cerclage removal group was 13.3 days and for the cerclage retention group, 9.4 days. Since an important question is whether retention of the cerclage is more likely to achieve the 48 hours of latency needed for optimal antenatal corticosteroid benefit, this comparison was done and no difference was seen with 11/20 (55%) achieving 48 hours in the cerclage retention group and 18/31 (58%) in the removal group. The rate of chorioamnionitis in the removal group was not statistically different but numerically greater in the retention group at 41% vs 25% in the removal group ( Table 2 ). Similarly, in comparing maternal outcomes after removing twins from the analysis, the number of undelivered after 1 week was not statistically different (17/31 [55%] vs 9/18 [50%]), nor was chorioamnionitis (8/31 [26%] vs 7/18 [39%]) in the removal vs cerclage retention groups, respectively. The mean gestational age at delivery in the removal group was 200 days (28.6 weeks) and 198 days (28.3 weeks) for the retention group. Birthweights were similar between the 2 groups. Despite the removal group having a nonstatistically significant longer latency and mean gestational age, their mean neonatal intensive care unit stay (in days) was longer (but not statistically different): 43 days compared to 33 days for the retention group. Composite neonatal outcomes are listed in Table 3 and were also similar between the 2 groups.

Randomization results

Galyean. Cerclage removal vs retention after PPROM. Am J Obstet Gynecol 2014 .

Table 3

Preterm premature rupture of membranes cerclage neonatal outcomes

Variable	Group 1: removal n = 33	Group 2: retention n = 30	P value
Composite neonatal outcome
n	32	30
Yes	16 (50.0%)	17 (56.5%)	.9109
Fetal or neonatal death
n	33	30
Yes	4 (12.1%)	4 (13.33%)	.5195
Respiratory distress syndrome
n	32	27
Yes	15 (46.8%)	14 (51.9%)	.9157
Documented sepsis
n	32	27
Yes	9 (28.1%)	9 (33.3%)	.3306
Grade-3 or -4 IVH
n	32	26
Yes	4 (12.5%)	4 (14.8%)	.5880
Necrotizing enterocolitis
n	32	27
Yes	2 (6.2%)	0 (0.0%)	.4950 a
NICU stay, d
n	32	26
Mean	43.0	33.4	.4835
Birthweight, g
n	33	30
Mean	1193.2	1180.8	.6203
Gestational age at delivery, d
n	33	30
Mean	200.0	198.1	.6771 b

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Abdominal wall endometriosis: 12 years of experience at a large academic institution
2. Triploid pregnancies: genetic and clinical features of 158 cases
3. Recurrent pregnancy loss: a risk factor for long-term maternal atherosclerotic morbidity?
4. Indocyanine green for intraoperative localization of ureter

Stay updated, free articles. Join our Telegram channel

Join