OBJECTIVES: Prophylactic antibiotics are recommended routinely for preterm prelabor rupture of membranes (PPROM), but there is an abundance of potential treatments and a paucity of comparative information. The aims of this network meta-analysis were to compare the efficiency of different antibiotic regimens on perinatal outcomes and to assess the quality of the current evidence.
METHODS: This was a network meta-analysis of randomized controlled trials comparing prophylactic antibiotics, or regimens of antibiotics, with each other or with placebo/no treatment, in women with PPROM. MEDLINE, Scopus, Cochrane Central Register of Controlled Trials, US Registry of Clinical Trials ( www.ClinicalTrials.gov) and gray literature sources were searched. The primary outcomes were neonatal mortality and chorioamnionitis; secondary outcomes included other measures of perinatal morbidity. Relative effect sizes were estimated using risk ratios (RR) and the relative ranking of the interventions was obtained using cumulative ranking curves. The quality of evidence for the primary outcomes was assessed according to GRADE guidelines, adapted for network meta-analysis.
RESULTS: The analysis included 20 studies (7169 participants randomized to 15 therapeutic regimens). For the outcome of chorioamnionitis, clindamycin + gentamycin (network RR, 0.19 (95% CI, 0.05-0.83)), penicillin (RR, 0.31 (95% CI, 0.16-0.6)), ampicillin/sulbactam + amoxicillin/clavulanic acid (RR, 0.32 (95% CI, 0.12-0.92)), ampicillin (RR, 0.52 (95% CI, 0.34-0.81)) and erythromycin + ampicillin + amoxicillin (RR, 0.71 (95% CI, 0.55-0.92)) were superior to placebo/no treatment. Erythromycin was the only effective drug for neonatal sepsis (RR, 0.74 (95% CI, 0.56-0.97)). Clindamycin + gentamycin (RR, 0.32 (95% CI, 0.11-0.89)) and erythromycin + ampicillin + amoxicillin (RR, 0.83 (95% CI, 0.69-0.99)) were the only effective regimens for respiratory distress syndrome, whereas ampicillin (RR, 0.42 (95% CI, 0.20-0.92)) and penicillin (RR, 0.49 (95% CI, 0.25-0.96)) were effective in reducing the rates of Grade-3/4 intraventricular hemorrhage. None of the antibiotics appeared significantly more effective than placebo/no treatment in reducing the rates of neonatal death, perinatal death and necrotizing enterocolitis. No network RR could be estimated for neonatal intensive care unit admission. The overall quality of the evidence, according to GRADE guidelines, was moderate to very low, depending on the outcome and comparison.
CONCLUSIONS: Several antibiotics appear to be more effective than placebo/no treatment in reducing the rate of chorioamnionitis after PPROM. However, none of them is clearly and consistently superior compared to other antibiotics, and most are not superior to placebo/no treatment for other outcomes. The overall quality of the evidence is low and needs to be updated, as microbial resistance may have emerged for some antibiotics, while others are underrepresented in the existing evidence. Copyright © 2019 ISUOG. Published by John Wiley & Sons Ltd.