Study designs of evaluations included in the review
Randomised controlled trials (RCTs), published as full papers or conference abstracts, were included if they reported the following comparisons: the induction of labour using PG versus expectant waiting (8 to 24 hours); and the induction of labour using PG versus induction with oxytocin.
Specific interventions included in the review
Trials of PG given vaginally, as either gel or pessaries, were included. In every case, the preparation used was natural PGE2. Evaluations of PG administered by general or cervical routes, or preparations of PG other than natural PGE2, were excluded. The PG regimens varied across the trials and included the following:
1 mg given as an initial dose and repeated 6 hours later;
1 mg given as an initial dose and repeated 6, 12 and 18 hours later;
1 mg given as an initial dose, followed by 3 mg given 1 hour and 4 hours later;
2 mg as an initial dose followed by 1 mg 6 hours later;
3 mg given as single dose;
two 3 mg doses given 4 hours apart;
two 3 mg doses given 6 hours apart; and
4 mg given as a single dose.
In all the included trials, it was stated that those patients that initially received PG were later given oxytocin if induction by PG failed. The interval between PG failure and oxytocin administration varied between the trials (range: 8 to 24 hours). In patients receiving expectant management, the interval between rupture of membranes and the induction of labour was within the range 8 to 24 hours, where the spontaneous onset of labour had not occurred. In patients receiving oxytocin, the starting dose by infusion varied between 1 and 3 mUI/minute, and increased by 1 to 2.5 mUI/minute every 10 to 30 minutes. The maximum doses, where reported, were either 24 or 56 mUI/minute, depending on the trial. Two trials described the use of antibiotics in cases where the time between the rupture of the membranes and the onset of labour exceeded 12 or 24 hours.
Participants included in the review
Trials recruiting women with premature rupture of membranes after 34 weeks' gestation were included. Cervical status was assessed using either the Bishop's score or by measuring cervical dilatation. The Bishop's score was determined by vaginal examination, whilst cervical dilatation was determined by either vaginal dilatation or the use of a speculum. All of the included trials recruited women with a singleton pregnancy, cephalic presentation, no previous Caesarean section, and a due date calculated according to the last menstrual period or early ultrasound scan. All of the trials excluded women with the following:
uterine contractions indicating the beginning of labour;
contraindications of induction, due to disproportion between the foetus and pelvis, diabetes, foetal distress, non-cephalic presentation, multiple birth, delayed growth, or haemorrhage;
the possibility of infection; and
maternal illness requiring immediate delivery.
Outcomes assessed in the review
There were no specific selection criteria relating to the outcomes of the primary studies. The following outcomes were assessed in the included trials:
the risk of maternal or foetal infection;
maternal fever peripartum and/or postpartum, defined as a temperature above 37.5 degrees C or 38 degrees C, depending on the trial;
delivery within 18 to 24 hours following induction;
recourse to administration of oxytocin;
delivery by Caesarean section;
instrumental delivery;
transfer to neonatal intensive care unit (ICU);
neonatal infection;
incidence of hypertonic uterus and ruptured uterus.
How were decisions on the relevance of primary studies made?
The authors do not state how the papers were selected for the review, or how many of the reviewers performed the selection.