Progestogens for preterm birth prevention: A systematic review and meta-analysis

OBJECTIVE: We systematically reviewed the effectiveness of progestogens for prevention of preterm birth among women with prior spontaneous preterm birth, multiple gestations, preterm labor, short cervix, or other indications. DATA SOURCES: We searched MEDLINE and EMBASE databases for English language articles published from January 1966 to October 2011. METHODS OF STUDY SELECTION: We excluded publications that were not randomized controlled trials or had fewer than 20 participants, identifying 34 publications, of which 19 contained data for Bayesian meta-analysis. TABULATION, INTEGRATION, AND RESULTS: Two reviewers independently extracted data and assigned overall quality ratings based on predetermined criteria. Among women with prior preterm birth and a singleton pregnancy (five randomized controlled trials), progestogen treatment decreased the median risk of preterm birth by 22% (relative risk [RR] 0.78, 95% Bayesian credible interval 0.68-0.88) and neonatal death by 42% (RR 0.58, 95% Bayesian credible interval 0.27-0.98). The evidence suggests progestogen treatment does not prevent prematurity (RR 1.02, 95% Bayesian credible interval 0.87-1.17) or neonatal death (RR 1.44, 95% Bayesian credible interval 0.46-3.18) in multiple gestations. Limited evidence suggests progestogen treatment may prevent prematurity in women with preterm labor (RR 0.62, 95% Bayesian credible interval 0.47-0.79) and short cervix (RR 0.52, 95% Bayesian credible interval 0.36-0.70). Across indications, evidence about maternal, fetal, or neonatal health outcomes, other than reducing preterm birth and neonatal mortality, is inconsistent, insufficient, or absent. CONCLUSION: Progestogens prevent preterm birth when used in singleton pregnancies for women with a prior preterm birth. In contrast, evidence suggests lack of effectiveness for multiple gestations. Evidence supporting all other uses is insufficient to guide clinical care. Overall, clinicians and patients lack longer-term information to understand whether intervention has the ultimately desired outcome of preventing morbidity and promoting normal childhood development.