Bacterial vaginosis (BV) is a recurrent disease in women despite treatment by antibiotics. This study investigated the impact of a vaginal probiotic, Lactobacillus crispatus IP174178* (Lc), on the rate of recurrence and time to recurrence. A prospective, multi-centre, double blind, randomised phase III trial in women with at least two documented episodes of BV in the previous year (diagnosis confirmed by presence of three Amsel criteria and a Nugent score≥7), and who had been clinically cured (i.e., no Amsel criteria) after oral metronidazole treatment (1g/day×7 days). The patients were randomised to receive vaginal capsules of either Lc or placebo, once a day, for 14 days over the first two menstrual cycles and another 14 days of the same treatment for the following two menstrual cycles. The primary efficacy endpoint was the number of patients with at least one bacteriologically confirmed recurrence of BV. Out of 98 assessable patients (mean age 35.7 years), 78 women were evaluated (20 patients had missing data). During the treatment period, 16/39 patients (41%) had at least one recurrence in the placebo group versus 8/39 patients (20.5%) in the Lc group (P=0.0497). The time to recurrence was longer by 28% in the Lc group (3.75±0.16 months) vs. the placebo group (2.93±0.18 months) (P=0.0298). Tolerability and safety were good in both groups. In women with recurrent BV after antibiotics, treatment with Lc IP 174178 administered over four menstrual cycles, could significantly reduce the rate of recurrence and increase the time to recurrence.