Ten RCTs (3,969 patients) were included.
Nine of the 10 RCTs reported that the treatment groups were similar at baseline, and nine described an assessment of compliance. Two RCTs excluded patients after randomisation. The results were presented for between 85.1 and 100% of the patients randomised.
Women without pre-term labour; pre-term delivery defined as delivery before 37 weeks.
The meta-analysis found no statistically significant difference between antibiotics and placebo in pre-term delivery for all patients combined or for high-risk patients. Using a random-effects model, the OR was 0.83 (95% CI: 0.57, 1.21) for all patients and 0.50 (95% CI: 0.22, 1.12) for high-risk patients (4 RCTs, 582 high-risk patients). Significant heterogeneity was detected for both meta-analyses (P<0.02 for both).
The meta-analysis found that oral antibiotics given for at least 7 days significantly reduced pre-term delivery compared with placebo (the abstract stated that these results apply to high-risk women, but table III did not specify the group). The OR (2 RCTS, 338 women) was 0.42 (95% CI: 0.27, 0.67). No significant heterogeneity was detected (P=0.38). For low-risk women, the meta-analysis found no statistically significant difference between antibiotics and placebo in pre-term delivery. The OR (3 RCTs, 2,530 women) was 1.25 (95% CI: 0.86, 1.81). No significant heterogeneity was detected (P=0.74).
Women with pre-term labour (1 RCTs, 25 women).
One small RCT found no statistically significant difference between antibiotics and placebo in pre-term delivery before 37 weeks. The OR was 0.31 (95% CI: 0.03, 3.24).
The results for other subgroup analyses were reported in the paper.