This updated review included 7 new RCTs (number of participants unclear as one trial did not report them) and the 7 RCTs (n=3,860) identified and included in the initial review.
Based on the Jadad scale, one trial scored 2 out of 5, three scored 3, eight scored 4 and two scored 5. No publication bias was detected.
The main results are reported below; further analyses are detailed in the original publication. Sensitivity analyses were carried out, excluding those studies with low internal validity, but this did not alter any of the overall estimates of treatment effect.
Low-risk women: 3 trials reported the results of treatment in low-risk women. No significant effect of treatment on delivery before 37 weeks was found (ARR -0.019, 95% CI: -0.056, 0.018). No significant heterogeneity was noted in these studies.
Average-risk women (general population): 8 trials reported the effect of treatment on delivery before 37 weeks and found no significant benefit (ARR 0.006, 95% CI: -0.009, 0.022). Three trials assessed the impact of treatment on delivery before 32 weeks and found no significant benefit (ARR 0.001, 95% CI: -0.008, 0.010). Seven trials found no benefit of treatment for low birth weight (ARR 0.000, 95%CI: -0.018, 0.018). No significant heterogeneity was noted in any of these analyses.
High-risk women: studies reporting the results of treatment in high-risk women were more heterogeneous overall and less statistical pooling was carried out. Studies reporting on the impact of treatment on delivery before 37 weeks were inconsistent: three showed some benefit, one indicated harm and the fifth found no benefit. Pooled data for 5 RCTs on the outcome of delivery before 34 weeks indicated no significant benefit of treatment (ARR 0.006, 95% CI: -0.067, 0.079).
An outcomes table, which summarised the clinical effects of screening and treating asymptomatic women for bacterial vaginosis, was presented in the main publication. Overall, some high-risk women may benefit from such an intervention, but many would experience either no benefit or may potentially be harmed.