Evidence from one randomized controlled trial (RCT) indicates that suppressive treatment with twice-weekly metronidazole vaginal gel may reduce rates of recurrence of bacterial vaginosis (BV), but it is associated with a higher risk of vaginal candidiasis.

Several small studies suggest that treatment with Lactobacillus species (either topically or orally) may reduce rates of recurrence, but dropout rates were high. More evidence is needed to confirm these effects and to determine the most appropriate strain of lactobacillus and the optimum dosing regimen.

One case series and one small RCT suggest that treatment with acetic acid gel or lactic acid gel may reduce the rate of recurrence of BV; however, there is insufficient evidence to make a recommendation regarding the use of vaginal acidification for recurrent BV.

Suppressive treatment with intravaginal metronidazole:

• One RCT investigated the efficacy of suppressive treatment with metronidazole vaginal gel to prevent recurrent BV [Sobel et al, 2006]:
• Women with recurrent symptomatic episodes of BV were initially treated with 0.75% metronidazole gel, 5 g at bedtime for 10 days. Women who had resolution of symptoms 3–5 days after completion of treatment were then randomized to receive 0.75% metronidazole gel (5 g) or placebo (5 g) twice a week on nonconsecutive days for 16 weeks, with a 12-week follow up without treatment:
• Of the 127 women initially evaluated, 112 (88%) responded to initial treatment and went on to be randomized. During the 16 weeks of the study, 17 (15%) of women were lost to follow up.
• By the end of the 16 weeks of suppressive treatment, fewer women receiving metronidazole experienced a recurrence of BV (13/51 [25%] with metronidazole vs. 26/44 [59%] with placebo; relative risk [RR] 0.43, 95% CI 0.25 to 0.73; p = 0.001).
• After the additional 12-week follow up (28 weeks in total), the difference remained significant (26/51 [51.0%] with metronidazole vs. 33/44 [75%] with placebo; RR 0.68, 95% CI 0.49 to 0.93; p = 0.02).
• Adverse effects were uncommon, but secondary vaginal candidiasis was more common in women using twice-weekly metronidazole gel. At some time during follow up, 59.3% of women randomized to metronidazole received antifungal treatment for vaginal candidiasis, compared with 34% of those randomized to placebo (p = 0.01).

Use of lactobacillus:

• CKS identified two RCTs investigating the efficacy of Lactobacillus species for recurrent BV:
• The first RCT was a cross-over study comparing the ingestion of yoghurt containing live L. acidophilus with pasteurized yoghurt in 46 women with recurrent BV (n = 20), recurrent candidiasis (n = 18), or both (n = 8) [Shalev et al, 1996]:
• Among women with recurrent BV, episodes of BV were significantly reduced in those ingesting live yoghurt (from 60% at the start of treatment to 25% after 1 month) compared with those ingesting pasteurized yoghurt (from 60% to 50%; p = 0.004). However, these results must be interpreted with caution because the dropout rate was very high and only seven women completed the entire treatment protocol.
• A second RCT compared the use of vaginal capsules containing a mixture of L. gasseri and L. rhamnosus with placebo vaginal capsules [Larsson et al, 2008]. After initial treatment with clindamycin 2% intravaginal cream, 100 women with BV were randomized to receive vaginal gelatine capsules containing either freeze-dried lactobacilli or identical placebo:
• At the end of the study (six menstrual cycles), 65% (24/37) of the lactobacilli-treated women had no recurrence of BV, compared with 46% (18/39) of the placebo-treated women. The difference between the groups in time from cure to relapse was statistically significant (p = 0.027) in favour of the lactobacilli treatment.
• A third study investigated the effectiveness of vaginal administration of L. rhamnosus after conventional metronidazole therapy [Marcone et al, 2008]. However, randomization in this study was not blinded, and there was no comparator arm. This study is therefore not considered further.

Acetic acid gel or lactic acid gel:

Intravaginal acetic acid gel (Aci-Jel^®) is no longer available in the UK. CKS found no other intravaginal acetic acid preparations available in the UK, although several intravaginal lactic acid preparations are available over-the-counter.

• CKS identified one case series investigating the efficacy of maintenance acetic acid vaginal gel (Aci-Jel^®) in the management of recurrent bacterial vaginosis [Wilson et al, 2005]. A total of 61 women identified as having proven recurrent BV were given standard treatment of either oral metronidazole, intravaginal metronidazole, or intravaginal clindamycin and assessed a week later. If on re-assessment vaginal flora was still abnormal (intermediate flora or BV) they were given further courses of treatment until flora became normal. When vaginal flora were normal, the women were advised to use acetic acid gel (2.5 g at night) for up to 3 months. A total of 49 women continued to use the acetic acid gel either nightly or for three nights after a recognized trigger factor. Mean follow up was 18 months.
• Recurrence was confirmed in 25 women (51%), with over half of these women developing recurrence within the first 3 months.
• The total number of recurrent episodes of BV in the 25 women was 44, with a mean of 1.8 per woman. Before using acetic acid gel, the mean recurrence rate in the 49 women was 4.4 per woman per year, with acetic acid gel this reduced to 0.6 per woman per year.
• CKS identified one small double-blind RCT comparing lactic acid vaginal gel (Lactal^®) with placebo in 42 women with recurrent BV [Andersch et al, 1990]. After initial treatment with 5 mL lactic acid gel once a day for 7 days, women were randomized to receive prophylactic treatment for 3 days each month for 6 months with lactic acid gel or placebo gel.
• Significantly more women treated with lactic acid gel showed clinical improvement compared with placebo (88% vs. 10%; p < 0.001).
• Lactobacilli was re-established in 83% of the lactic acid group and 16% of the placebo group.