Bacterial vaginosis (BV) is one of the most frequent female lower genital tract infections, not only in pregnancy but throughout the reproductive life. Studies from Europe and the USA have found prevalence between 4.9% and 36.0% [¹]. The first signs of BV are radical changes in the vaginal ecosystem. H₂0_2-producing lactobacilli, which are present in 96% of women with normal vaginal bacterial flora, are markedly reduced or lost, while microorganisms like Gardnerella vaginalis and obligate anaerobes prevail [²]. The cause of this change is not clear [³] and the microorganisms responsible for the shift in the flora have still to be identified [⁴]. BV may be due not only to the excessive bacterial growth, but also to the formation of a dense bacterial biofilm adherent to the vaginal mucosa.

The biofilm formed by Gardnerellavaginalis in BV was first identified by electron microscopy as a dense tissue strongly adherent to the vaginal epithelium, and made up of bacterial cells packed inside a network of polysaccharide fibrils [⁵, ⁶]. Later, Swidsinki et al., investigating vaginal biopsies by bacterial rDNA fluorescent in situ hybridization, suggested that the bacterial biofilm played a primary role in the development of BV [⁷].

Costerton et al. and Swidsinki et al. found a dense bacterial biofilm, coating at least half the epithelial surface, in 90% of biopsies from women with BV, and in only 10% of healthy women [⁷, ⁸]. The presence of the biofilm enables the bacterial cells to reach higher concentrations (up to 10¹¹ bacteria/mL) than in vaginal fluid and boosts their resistance to both the host immune system and the antimicrobials [⁹, ¹⁰]. In fact, the drugs hardly reach the bacteria, residing inside the film in a quiescent state, leading to an up to 1000-fold antimicrobial decreased activity [⁹, ¹¹]. This observation might provide an explanation of the high rates of BV relapses [¹⁰, ¹²].

BV has aroused interest in the last few years being considered as a predisposing factor for HIV, Type II Herpes symplex virus, Chlamydia trachomatis infections, as well as for trichomoniasis and gonorrhea [¹³, ¹⁴]; BV can be also a cause for complications like late abortion [¹⁵], premature rupture of the amniotic membrane [¹⁶], chorio-amnionitis [¹⁷], post-partum endometritis [¹⁸, ¹⁹, ²⁰], and failure of in vitro fertilization and embryo transfer [¹³, ¹⁴]. Particular attention has been recently paid to Atopobium vaginae, a newly identified bacterium, belonging to the Coriobacteriaceae family, which is believed to be at least a partial cause of the above mentioned complications [¹³]. The genus Atopobium, described for the first time in 1992, includes bacteria previously classified as lactobacilli. Rodriguez first identified A. vaginae in a study on vaginal lactobacilli [²¹]. A. vaginae 16s rRNA gene differs from the other species belonging to Atopobium genus by approximately 3-8% [²², ²³]; this enabled Rodriguez to identify it as a new species. The isolate can be distinguished from A. minutum, A. parvulum, and A. rimae by biochemical tests and protein electrophoresis of the whole cell (Table 1). Gram stain shows A. vaginae as a small coccus, rounded or oval, or rods, visible as single cells, in pairs or in short chains (Fig. 1).

This aerobic facultative, gram-positive bacterium cannot be easily isolated by classical microbiological methods [¹⁴, ²⁴]. It is hardly detected in healthy women vaginal fluid but is commonly found in the vagina of patients with BV: 50% according to Burton [²⁵, ²⁶], 70% according to Ferris [²⁷], and more than 95% according to Verhelst et al. [²⁴] and Verstraelen et al. [²⁸]. In symptomatic BV it has been detected together with Gardnerella vaginalis in the biofilm adherent to the vaginal mucosa [²⁴]. This was confirmed by Swidsinski et al. [⁷] who, by examining the composition and structural organization of the biofilm, found that Gardnerella vaginalis accounted for 60-95% of the film mass. In addition, in 70% of bioptic samples, Atopobium vaginae accounted for the 1-40% of the film mass. Lactobacillus concentrations were lower than 10⁶ CFU/mL, making up only 5% of the biofilm (Fig. 2).

Concerning the pharmacological therapy, CDC recommends either oral or topical (vaginal gel) metronidazole once a day for 5 days as first choice for BV. Efficacy is comparable to topical clindamycin [²⁹]. Cure rates, following intravaginal treatment with metronidazole or clindamycin, account for 80-90% at the end of treatment and one month after the end of therapy [¹³, ¹⁴, ³⁰]. However, three months after the end of therapy the rate of relapses can overcome 30%. Persistence of an adherent bacterial biofilm, containing mostly G. vaginalis and A. vaginae, seems to be the main reason for failure of BV treatment [³⁰]. Suppressive treatment with metronidazole gel and physiological approaches (use of probiotics or acidifying) have been investigated with variable results [³¹]. Moreover, long-term treatment with metronidazole is not recommended because of the high incidence of gastrointestinal adverse reactions, the risk of peripheral neuropathy, and Candida super infection [³²].

Failures with metronidazole in patients with recurrent or persistent BV [³³, ³⁴] might conceivably reflect the newly found mechanism of formation of a biofilm containing G. vaginalis together with A. vaginae [⁷, ⁹, ¹³, ²⁸] (Fig. 3). The fact that A. vaginae is resistant to metronidazole, and that the bacterium creates a biofilm in which it is associated with G. vaginalis, complicates the response to the antibiotic [⁹, ¹³, ²⁸]. Though clindamycin is more active than metronidazole against both G. vaginalis and A. vaginae, its negative effects on lactobacilli leave the way open to microbial disorders that can cause frequent super infections and recurrences. Moreover, an increasing resistance to antibiotics that act like clindamycin, by blocking protein synthesis has been reported. A randomized prospective trial compared 119 women assigned to two therapeutic regimens for BV: either metronidazole vaginal gel for five days, or clindamycin vaginal tablets for three days. The clinical efficacy was comparable in the two arms: after 7-12 days about 80% of the patients were cured, but this percentage fell down to about 50% after 35-45 days. Following clindamycin treatment – but not metronidazole - there was a steep rise in the percentage of women with at least one clindamycin resistant strain isolated. Moreover, 70-90 days after the end of treatment, about 80% of the women who received clindamycin presented in their vaginal swabs anaerobic bacteria resistant to that drug [³⁵].

Togni et al. [³⁶] compared the in vitro susceptibility of A. vaginae to nifuratel, metronidazole and clindamycin. Susceptibility to metronidazole was variable, with MIC ranging from 8 to 256 µg/mL. Nifuratel and clindamycin inhibited the growth of all the tested strains, with MIC from 0.125 to 1 µg/mL and below 0.125 µg/mL, respectively (Table 2). The findings related to metronidazole and clindamycin are in line with previously published studies [³⁷].

In the same study, the activity of these antibiotics was assayed on lactobacilli and G. vaginalis. Either nifuratel and metronidazole did not affect the normal lactobacterial flora, while clindamycin inhibited all tested strains of lactobacilli. Nifuratel and metronidazole were both highly active against G. vaginalis (Fig. 4). The susceptibility of Atopobium vaginae to metronidazole and clindamycin, and the action on lactobacilli and G. vaginalis were in line with previous reports [^37-39]. To summarise, nifuratel was active against A. vaginae and G. vaginalis strains without affecting lactobacilli; metronidazole was active against A. vaginae, but only at very high concentrations, partially active against G. vaginalis, and did not affect lactobacilli; clindamycin was extremely effective against A. vaginae and G. vaginalis, but it also affected the lactobacilli, altering the vaginal ecosystem.