Objective: Efficacy study of suppressive vaginal metronidazole in reducing recurrent symptomatic episodes of bacterial vaginosis. Study design: Multicenter prospective study with initial 10-day open-label metronidazole gel in which asymptomatic responders randomly assigned to receive twice weekly metronidazole vaginal gel or placebo for 16 weeks and off therapy for 12 weeks. Results: Of 157 eligible women with recurrent bacterial vaginosis, 112 of 127 returning evaluable women (88.2%) responded clinically and were randomly assigned. During suppressive therapy, recurrent bacterial vaginosis occurred in 13 women (25.5%) receiving metronidazole and 26 (59.1%) receiving placebo (MITT analysis, relative risk [RR] 0.43, CI = 0.25-0.73, P = .001). During the entire 28-week follow-up, recurrence occurred in 26 (51.0%) on treatment compared with 33 (75%) on placebo (RR 0.68, CI = 0.49-0.93, P = .02). Probability for remaining cured was 70% for metronidazole compared with 39% on placebo, which declined to 34% and 18%, respectively, by 28 weeks follow-up. Adverse effects were uncommon; however, secondary vaginal candidiasis occurred significantly more often in metronidazole-treated women (P = .02). Conclusion: Suppressive therapy with twice-weekly metronidazole gel achieves a significant reduction in the recurrence rate of bacterial vaginosis; however, secondary vaginal candidiasis is common. © 2006 Mosby, Inc. All rights reserved.