Two RCTs (sample size reported as n=160 in the forest plot and n=187 in the text) were included in the meta-analysis. One trial was placebo-controlled. Follow-up was to the end of pregnancy. Jadad scores were 3 and 5.
Pooled results showed a statistically significant risk reduction for miscarriage (RR 0.48, 95% CI 0.25 to 0.92, I2 = 23%; two trials). Relative reductions in miscarriage rates were 36% and 75%. One trial (n=115) showed a statistically significant risk reduction for pre-term birth (RR 0.31, 95% CI 0.11 to 0.90).
No reported safety concerns arose from the treatment.