Preventing postpartum hemorrhage after cesarean delivery: a network meta-analysis of available pharmacologic agents

Abstract

Background: Postpartum hemorrhage causes a quarter of global maternal deaths. The World Health Organization recommends oxytocin as the first line agent to prevent hemorrhage during cesarean delivery. However, some randomized controlled trials suggest that other uterotonics are superior.

Objective: We conducted a network meta-analysis comparing the ability of pharmacologic agents to reduce blood loss and minimize the need for additional uterotonics during cesarean delivery.

Data sources: We searched the Cochrane Central Register of Controlled Trials, Embase, and MEDLINE databases from inception to May 2020.

Study eligibility criteria: We included randomized controlled trials that compared oxytocin, carbetocin, misoprostol, ergometrine, carboprost, or combinations of these in the prevention of postpartum hemorrhage during cesarean delivery.

Methods: We performed a systematic review followed by an NMA in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. Quality of the evidence was assessed with the Confidence in Network Meta-Analysis approach and Grading of Recommendations, Assessment, Development and Evaluations tool within the summary of findings table. Our primary outcomes were the estimated blood loss and need for additional uterotonics. Secondary outcomes included nausea and postpartum hemorrhage of >1000 mL. We performed sensitivity analyses to explore the influence of surgical context and oxytocin administration strategy.

Results: A total of 46 studies with 7368 participants were included. Of those, 21 trials (6 agents and 3665 participants) formed the "estimated blood loss" network and, considering the treatment effects, certainty in the evidence, and surface under the cumulative ranking curve scores, carbetocin was assessed to probably be superior to oxytocin, but only in reducing the estimated blood loss by a clinically insignificant volume (54.83 mL; 95% confidence interval, 26.48-143.78). Misoprostol, ergometrine, and the combination of oxytocin and ergometrine were assessed to probably be inferior, whereas the combination of oxytocin and misoprostol was assessed to definitely be inferior to oxytocin. A total of 37 trials (8 agents and 6193 participants) formed the "additional uterotonic" network and, again, carbetocin was assessed to probably be superior to oxytocin, requiring additional uterotonics 185 (95% confidence interval, 130-218) fewer times per 1000 cases. Oxytocin plus misoprostol, oxytocin plus ergometrine, and misoprostol were assessed to probably be inferior, whereas carboprost, ergometrine, and the placebo were definitely inferior to oxytocin. For both primary outcomes, oxytocin administration strategies had a higher probability of being the best uterotonic, if initiated as a bolus.

Conclusion: Carbetocin is probably the most effective agent in reducing blood loss and the need for additional uterotonics. Oxytocin appears to be more effective when initiated as a bolus.

Keywords: CD; NMA; PPH; carbetocin; carboprost; ergometrine; misoprostol; oxytocin; prostaglandin.