386: A network meta-analysis of cervical ripening interventions


      Prior studies and reviews of cervical ripening interventions to initiate labor induction focused on pair-wise comparisons of interventions with regard to select outcomes, making it difficult to establish an overall hierarchy of ideal approaches. We synthesized evidence from a systematic review into a network of 12 cervical ripening interventions and 7 outcomes.

      Study Design

      We searched the Cochrane Library and MEDLINE for randomized controlled trials of cervical ripening interventions. Citations were screened in duplicate with discrepancies adjudicated independently. Eligible interventions included vaginal (gel and tablet), oral, and intracervical prostaglandins; amniotomy; vaginal, oral, and buccal/sublingual misoprostol; balloon catheter; and membrane sweeping. Eligible outcomes included vaginal delivery after 24 hours, instrumental delivery, cesarean delivery, uterine hyperstimulation, Apgar score, postpartum hemorrhage, and neonatal intensive care unit admission. A network meta-analysis was conducted in a Bayesian setting, with models for each outcome specified as generalized linear mixed models to account for patient clustering within studies.


      312 articles were eligible. Across the 7 outcomes, the number of interventions with data ranged from 8 to 12, and the number of studies and participants, respectively, ranged from 38 and 7,492 (vaginal delivery > 24 hours) to 257 and 47,287 (cesarean delivery). Using cesarean delivery as an example (network shown in Figure), buccal/sublingual misoprostol and oral misoprostol were among the least likely to result in cesarean delivery while amniotomy and placebo were among the most likely. The favored interventions differed somewhat across outcomes. For example, buccal/sublingual misoprostol was among the most effective interventions for all outcomes except hyperstimulation and balloon catheter was favored for all outcomes except achieving vaginal delivery within 24 hours. Conversely, placebo had high probability of being among the least effective interventions for all outcomes except hyperstimulation.


      The ranking of optimal cervical ripening interventions based on findings of published trials differs by outcome, but buccal/sublingual misoprostol and balloon catheter were shown to be among the most likely interventions to have favorable outcomes.
      Figure thumbnail fx1