Eleven RCTs were included in the review (n= 6,558). Five trials were considered to be high quality, five trials low quality, and one was unclear.
Overall, no statistically significant difference was reported between treatment and control groups for preterm birth before 37 weeks gestation (OR 0.93, 95% CI 0.79 to 1.10, I2=61%; 11 trials).
There was no benefit with treatment in the subgroup analysis of high-quality trials (OR 1.15, 95% CI 0.95 to 1.40, I2=1%; five trials, n=4,718). There was a statistically significant benefit with treatment in low-quality trials (six trials, n=1,840).
Overall, there was no statistically significant difference between groups for any of the secondary outcomes. This result was unchanged in the subgroup analysis of high-quality trials. Only in low-quality trials was statistical significance reported for low birthweight and overall adverse pregnancy outcomes.
There was evidence of publication bias.