The trial was sponsored, designed, and conducted by the Obstetrix Collaborative Research Network, a consortium of maternal-fetal medicine practices across the United States. The trial was approved by the independent Institutional Review Board (IRB) at each site. The study was conducted under Investigational New Drug (IND) Number 107785 through the United States Food and Drug Administration (FDA). The trial was registered on http://clinicaltrials.gov, #NCT01119963. An independent Data and Safety Monitoring Board (DSMB) supervised the trial, reviewed adverse event reports, and approved the premature termination of the trial. Additional File 1 contains the complete, IRB-approved, final protocol.
Women were eligible if they were at least 18 years old, had a singleton pregnancy at 23.0 to 31.9 weeks of gestation, and PROM. PROM was defined as either (a) documentation of vaginal leakage of indigo carmine dye instilled via amniocentesis; (b) a positive Amnisure® test (Amnisure International, Cambridge, MA); or (c) two or more of the following: nitrazine test of vaginal secretions with pH 7 or higher, ferning of vaginal secretions, gross pooling of clear fluid in the posterior vaginal fornix, or ultrasound exam showing oligohydramnios. We excluded women with active preterm labor, defined as 8 or more uterine contractions per hour that were perceived by the patient and/or a cervical dilation 4 cm or more. The definition of PROM did not require that the rupture of membranes had occurred before the onset of labor (commonly called "premature" or "prelabor" PROM); that is, women were eligible if they had labored before PROM provided that they were not in active labor, as defined, at the time of enrollment. We excluded women with contraindications to expectant management (such as suspected intraamniotic infection, nonreassuring fetal heart rate tracing, fetal death, preeclampsia, active uterine bleeding, or documented fetal lung maturity), with known fetal abnormalities (such as major congenital malformation, viral infection, or hydrops), with history of allergy to 17P or castor oil, with medical conditions that might adversely interact with 17P (such as asthma requiring medications, renal insufficiency, seizure disorder, ischemic heart disease, cholecystitis, impaired liver function, or history of venous thromboembolism, breast cancer or depression requiring hospitalization), with medical conditions treated with systemic steroid medications, or with a cervical cerclage present at the time of PROM. Eligible women were approached by a physician or research nurse and were offered participation in the trial.
Women who gave informed consent were randomly assigned in a 1:1 ratio to receive either 17P (250 mg in castor oil, 1 mL total volume, intramuscular injection weekly) or an identical-appearing placebo (1 mL castor oil only). A computer-generated random-number sequence was used by the trial statistician to generate a randomization code book kept at each site's inpatient pharmacy. Study medications (17P versus placebo) were prepared by McGuff Pharmaceuticals (Newport, California) according to the IND specifications, which followed current Good Manufacturing Practices (cGMP). Medications were provided in 5 mL multi-dose vials, specially labeled for the study and number-coded to correspond with the randomization code book. Randomization was stratified by gestational age, 23.0-25.9 weeks, 26.0 to 28.9 weeks, and 29.0 to 31.9 weeks. Participants and research personnel were blinded to group assignment throughout, from before enrollment until after completion of all case report forms and resolution of all data queries. The code was not broken until after the database was "locked."
Each week from randomization until delivery, the assigned medication (17P or placebo) was drawn into a syringe labeled "Study medication: progesterone or placebo" by the inpatient pharmacy at each site, delivered to the nursing unit, and administered by a registered nurse. In the syringes, 17P and placebo appeared visually identical.
Other than the administration of study medication (17P or placebo), the remainder of each patient's clinical care followed standard clinical management for PROM. The investigators agreed in advance on many items of standard management. These included inpatient hospitalization of all patients until delivery, administration of a single course of antenatal corticosteroids, broad-spectrum antibiotic prophylaxis for 1 week, avoidance of tocolytics (except during the 48 h after the first dose of corticosteroids), and fetal heart rate monitoring at least 1 h daily. We encouraged amniocentesis to rule-out intraamniotic infection before enrollment, but did not require this. We discouraged digital examination of the cervix, but did not exclude subjects who had already had digital examinations. The protocol called for collection of amniotic fluid from vaginal leakage at 32.0 weeks or beyond for assessment of fetal lung maturity according to whatever test was in use at the local site. Intrapartum antibiotic prophylaxis (e.g. group B streptococcus coverage) was encouraged but not mandated. A protocol was in place for evaluation and management of suspected membrane resealing, but this was not used because all the participants continued leaking amniotic fluid until delivery. Management decisions not specifically addressed by protocol were left to the discretion of the managing physician, including route of delivery, choice of antibiotics, and other matters.
The primary outcome was defined as prolongation of the pregnancy until a favorable gestational age, which we defined as either 34.0 weeks of gestation or documentation of fetal lung maturity at 32.0 to 33.9 weeks. The investigators agreed that continuation of pregnancy beyond these time points was not indicated. Secondary outcomes were latency (interval from randomization to delivery) and composite neonatal morbidity, which was defined per Mercer et al.  as one or more of: stillbirth, neonatal death, infant death before hospital discharge, respiratory distress syndrome (RDS), intracranial hemorrhage (ICH) grade 3 or 4, necrotizing enterocolitis (NEC) stage 2 or 3, culture-proven neonatal sepsis within 72 h of birth. In addition to the components defined by Mercer et al. , our definition of composite neonatal morbidity included periventricular leukomalacia (characteristic lesions in the subcortical white matter seen on cerebral imaging studies within 96 h of birth.)
Sample size calculations indicated that we would need 105 participants in each group (17P or placebo) to yield 80% power to detect a 20% absolute increase in the rate of the primary outcome from 30% in the placebo group to 50% in the 17P group. This was adjusted to 111 per group to account for a possible 5% rate of loss-to-follow-up due to membrane resealing or other factors. A single interim analysis was planned when data were available for 50% of the total sample, but this was not performed owing to early termination of the study. Because of early termination, the trial is grossly underpowered; therefore, we present only descriptive statistics by treatment group and we did not perform any formal statistical tests of potential between-group differences.
The trial was designed, conducted, analyzed, and reported according to the principles outlined in the CONSORT Statement and Checklist .