Objective
We sought to determine whether prophylactic treatment with 17-alpha-hydroxyprogesterone caproate (17Pc) in twin pregnancy will reduce neonatal morbidity (primary outcome) by prolonging pregnancy (secondary outcome).
Study Design
This was a double-blind, randomized clinical trial. Mothers carrying dichorionic-diamniotic twins were randomly assigned (in a 2:1 ratio) to weekly injections of 250 mg of 17Pc or placebo, starting at 16-24 weeks and continued until 34 weeks.
Results
In all, 160 women were randomized to 17Pc and 80 to placebo. Composite neonatal morbidity occurred with similar frequency in the 17Pc and placebo groups (14% vs 12%, respectively, P = .62). Mean gestational age at delivery was not affected by 17Pc (35.3 vs 35.9 weeks, P = .10), but a 3-day difference in median gestational age favored placebo ( P = .02). There were no perinatal deaths with 17Pc and 3 with placebo.
Conclusion
In twin pregnancy, prophylactic treatment with 17Pc did not prolong gestation or reduce neonatal morbidity.
Twin pregnancy carries a risk of serious neonatal morbidity due to a high rate of preterm birth. In the United States, 57% of twins are born with low birthweight, compared to 6% of singletons. Infant mortality approaches 3% among twins, nearly 5 times higher than in singletons, primarily because of complications of prematurity.
Recent clinical trials have suggested that progestins may reduce the risk of preterm birth in certain clinical settings, such as a history of preterm birth, a short cervix, or after arrest of preterm labor. A Committee Opinion from the American College of Obstetrics and Gynecology (ACOG) concluded that use of progestins is beneficial for a “select group of very high risk women,” but that “further studies are needed to evaluate the use of progesterone in patients with other high-risk obstetric factors, such as multiple gestation… .” At the inception of the present trial, only a single study had investigated whether progestins might reduce prematurity in multifetal pregnancy. That study found no benefit of prophylactic 17-alpha-hydroxyprogesterone caproate (17Pc) in twins, but the drug was not started until a mean gestational age of 29 weeks.
We hypothesized that prophylactic 17Pc started in the second trimester in women with twin pregnancy might reduce the rate of preterm birth and might thereby reduce the risks of neonatal morbidity and mortality. The present study was undertaken to test this hypothesis.
Materials and Methods
The study was performed by the Obstetrix Collaborative Research Network, a consortium of maternal-fetal medicine practices across the United States. The participating practices and personnel are listed in the Appendix . In most cases, subjects were enrolled and followed up on an outpatient basis antenatally, with delivery and neonatal management at an appropriate hospital. The study was approved by the institutional review board at each site. The study was conducted under US Food and Drug Administration (FDA) Investigational New Drug number 69-536, which covered both the present study of twin pregnancy and our previously reported concurrent, parallel trial of 17Pc in triplet pregnancy. The 2 trials were registered together at www.clinicaltrials.gov , NCT00163020 , where it was explicitly stated that the 2 trials were considered separate as they differed in their primary inclusion criterion (twins vs triplets). An independent data and safety monitoring board supervised the trial, reviewed adverse event reports, and conducted an interim analysis of efficacy.
Women were eligible if they had a dichorionic-diamniotic twin pregnancy at 15-23 weeks of gestation and if they had completed a detailed ultrasound examination showing no major fetal anomalies. Women were excluded if they were <18 years old; had taken any progestins >15 weeks of gestation; or had symptomatic uterine contractions, rupture of the fetal membranes, any contraindication to prolonging the pregnancy, any preexisting condition that might be worsened by progesterone, or a preexisting medical condition carrying a high risk of preterm delivery. We did not exclude women with a history of preterm birth, women who had undergone multifetal pregnancy reduction, or those who had undergone prophylactic cervical cerclage.
Eligible subjects were offered preliminary enrollment. After informed consent, a trial injection of 1 mL of castor oil was given intramuscularly and the subject was asked to return for an enrollment-completion visit approximately 1 week later. Those who returned were randomly assigned to receive either 17Pc (250 mg in 1 mL castor oil) or identical-appearing placebo (1 mL castor oil) and were given the first dose of medication at that visit. Randomization was allowed at ≥16 weeks of gestation, but not >24 weeks.
Randomization followed a computer-generated scheme kept at McGuff Pharmaceuticals Inc (Santa Ana, CA), the central compounding pharmacy. The scheme called for 2 subjects to be assigned to 17Pc for every 1 subject assigned to placebo, stratified so that each study site would have a similar ratio. Progesterone or identical-appearing placebo was compounded by the pharmacy according to the Investigational New Drug specifications; shipped in advance to each study site in coded, prenumbered kits; and replenished as needed. To randomize a subject, the research nurse contacted the central pharmacy by telephone or fax to obtain the code number for the kit assigned to that patient. Subjects, physicians, and study personnel remained blinded as to group assignment until after completion of the trial. The assigned medication was repeated weekly until 34.0 weeks of gestation or delivery, whichever occurred first.
Patients were given the option to return to the study site for weekly repeat injections or to have a partner give the injections at home. Those who chose home injections were asked to bring their partner for training regarding medication handling, injection technique, and biohazards disposal. Subjects were instructed to keep an injection diary and complete a side-effects questionnaire for each injection. To assess compliance, we reviewed these diaries and also measured the unused medication returned by the patient at the conclusion of her dosing. Outcomes for each patient were tabulated according the assigned group (17Pc vs placebo) regardless of her compliance.
At 24-26 weeks of gestation, a cervicovaginal swab was obtained for fetal fibronectin testing and a transvaginal ultrasound was performed for cervical length assessment. These procedures were not part of routine clinical care of twins at some of the study sites and subjects at those sites were not charged for these assessments.
Other than the administration of 17Pc vs placebo and the visit at 24-26 weeks, investigators were encouraged to render standard clinical care for twins, as outlined in an ACOG Educational Bulletin on multiple gestations, a synopsis of which was incorporated into the protocol.
The primary study outcome was composite neonatal morbidity, defined as ≥1 of: perinatal death (stillbirth, neonatal death, miscarriage), respiratory distress syndrome, use of oxygen therapy at 28 days of life, neonatal sepsis proven by blood culture, pneumonia, intraventricular hemorrhage (grade 3 or 4), periventricular leukomalacia, necrotizing enterocolitis requiring surgery, retinopathy of prematurity, or newborn asphyxia (ischemic injury of vital organ with pH < 7.0 from umbilical artery or newborn artery). Secondary outcomes included the individual neonatal morbidities listed above, gestational age at delivery, birthweight, and maternal side effects.
Statistical analysis of the primary outcome (composite neonatal morbidity) and secondary neonatal outcomes used a repeated measures approach (generalized estimating equations for dichotomous variables with an exchangeable correlation structure and a mixed linear model for continuous variables using a compound symmetry correlation structure) where each baby within given pregnancy was considered the repeated measure. Differences between treatment groups in baseline maternal characteristics, and maternal delivery characteristics and complications were determined using a χ 2 test or Fisher’s exact test for dichotomous variables and a Wilcoxon rank sum test for continuous or ordinal variables. Time from randomization to delivery was analyzed using a staggered entry Kaplan-Meier analysis (adjusting for gestational age at randomization) and differences between the treatment groups were determined using the log rank test. All analyses were performed using software (SAS 9.2, SAS Institute Inc, Cary, NC). All statistical tests are reported with 2-sided P values.
We estimated that the primary outcome would occur in 45% of the placebo group based on estimates from the Pediatrix Medical Group neonatal database of 12,302 twin newborns from 124 centers nationwide. We calculated that a sample of 480 newborns (320 in the 17Pc group and 160 placebo) would yield a power of 80% at an alpha level of 0.05 to detect a reduction of neonatal morbidity to 30% in the 17Pc group, with an assumption of independence between newborns. To obtain this number of newborns, we planned to enroll 240 mothers (roughly 160 in the 17Pc group and 80 placebo).
Interim analyses of the primary outcome were planned upon completion of 50% and 75% of the case reports. Only the first of these was actually performed. By the time 75% of the patients had delivered and case report forms had been completed, all but 8 of the planned total of 240 subjects had been enrolled and the data and safety monitoring board determined that a second interim analysis would have been moot. To correct for the interim analysis, the alpha level for the primary outcome was adjusted to 0.0466 based on the O’Brien-Fleming spending function. For all other analyses, no adjustments were made and an alpha level of 0.05 was used.
Results
Recruitment occurred at 18 sites from November 2004 through August 2009. We identified 1450 women with twins, of whom 648 were eligible for trial inclusion ( Figure 1 ). Of these, 254 gave consent and were given a trial injection of placebo and 240 returned the following week for random treatment allocation, 160 of whom were assigned to 17Pc and 80 to placebo. Outcome data were available for all mothers and newborns in the 17Pc group. In the placebo group, 2 mothers and their 4 newborns were lost to follow-up shortly after randomization.