Objective: To ascertain whether metronidazole treatment of women with a heavy growth of Gardnerella vaginalis during mid-pregnancy would reduce the risk of spontaneous preterm birth.
Design: A multicentre, randomised, placebo-controlled trial.
Setting: Four metropolitan hospitals.
Participants: Eight hundred and seventy-nine singleton women with a heavy growth of G. vaginalis or Gram stain indicative of bacterial vaginosis at 19 weeks of gestation.
Interventions: Oral metronidazole (400 mg) or placebo twice daily for two days at 24 weeks of gestation, and at 29 weeks if G. vaginalis found in test-of-cure swab four weeks after treatment.
Main outcome measures: Spontaneous preterm birth less than 37 weeks.
Results: Intention-to-treat analysis showed no difference between metronidazole and placebo groups in overall preterm birth (31/429 [7.2%] vs 32/428 [7.5%]) or spontaneous preterm birth (20/429 [4.7%] vs 24/428 [5.6%]). Among the 480 women with bacterial vaginosis treatment had no effect on spontaneous preterm birth (11/242 [4.5%] vs 15/238 [6.3%]). In the subset of 46 women with a previous preterm birth, women in the metronidazole group showed a significant reduction in spontaneous preterm birth (2/22 [9.1%] vs 10/24 [41.7%], OR 0.14, 95% CI 0.01-0.84). A treatment effect was also found in compliant women with a previous preterm birth and bacterial vaginosis (0/14 [0%] vs 6/17 [35.3%], OR 0.0, 95% CI 0.0-0.94).
Conclusion: Metronidazole treatment of women with a heavy growth of G. vaginalis or bacterial vaginosis did not reduce the preterm birth rate. Among women with a previous preterm birth, treatment reduced the risk of spontaneous preterm birth. Further studies are required to confirm these findings.