Temporal variability of human vaginal bacteria and relationship with bacterial vaginosis

Abstract

Background: Little is known about short-term bacterial fluctuations in the human vagina. This study used PCR to assess the variability in concentrations of key vaginal bacteria in healthy women and the immediate response to antibiotic treatment in women with bacterial vaginosis (BV).

Methodology/principal findings: Twenty-two women assessed for BV using Amsel's criteria were evaluated daily for 7 or 14 days, then at 2, 3 and 4 weeks, using a panel of 11 bacterium-specific quantitative PCR assays. Participants with BV were treated with 5 days of intravaginal metronidazole. Participants without BV had vaginal biotas dominated by lactobacilli, whose levels fluctuated with menses. With onset of menstruation, quantities of Lactobacillus jensenii and Lactobacillus crispatus decreased and were found to be inversely related to Gardnerella vaginalis concentrations (p<0.001). Women with BV had a variety of fastidious bacteria whose concentrations dropped below detection thresholds 1-5 days after starting metronidazole. Recurrent BV was characterized by initial profound decreases of BV-associated bacteria after treatment followed by subsequent increases at relapse.

Conclusions/significance: The microbiota of the human vagina can be highly dynamic. Healthy women are colonized with Lactobacillus species, but levels can change dramatically over a month. Marked increases in G. vaginalis were observed during menses. Participants with BV have diverse communities of fastidious bacteria that are depleted by vaginal metronidazole therapy. Women with recurrent BV initially respond to antibiotic treatment with steep declines in bacterial concentrations, but these bacteria later reemerge, suggesting that antibiotic resistance in these bacteria is not an important factor mediating BV recurrence.

Figure 1. Variability of vaginal bacteria in women without BV.

The line graphs show the changes in vaginal bacteria measured as 16S rRNA gene copies per swab in a participant exhibiting a stable bacterial biota (A) and another with dynamic patterns (B). The red line indicates menses. Clinical diagnosis was performed using Amsel's criteria (green box in upper graph indicating negative for BV). Gram staining using the Nugent criteria (the score is indicated as a number from 0–10 in the box) was also performed. A score of 0–3 indicates negative for BV, 4–6 reflects intermediate BV flora and 7–10 denotes positive for BV. Levels of the human 18S rRNA gene (red circles) were used to assess the amount of vaginal fluid loaded on each swab as reflected by human cell content. BVAB denotes bacterial vaginosis associated bacterium.

Figure 2. Variability of Gardnerella vaginalis during menses.

The line graphs show increases in G. vaginalis levels during menses while concentrations of L. crispatus and L. jensenii decreased. G. vaginalis quantities declined below detection thresholds post-menses with simultaneous increases in L. jensenii and L. crispatus levels. Participant C (Figure 2A) was not colonized with L. iners while Participant D (Figure 2B) was. Levels of L. iners were observed to increase with G. vaginalis concentrations and subsequently declined post-menses. Both participants were negative for BV by Amsel's criteria (green boxes) and had Nugent scores of 0 at the entry and follow up visits.

Figure 3. Eradication of BV bacteria post-antibiotic treatment.

Participant E (Figure 3A) and Participant F (Figure 3B) are two women with BV by Amsel criteria (red boxes) and by Nugent criteria (Score = 8) at study entry. There was a difference in the rate of decrease of BV bacteria post-metronidazole therapy. Participant E was cured of BV at the 1 month visit as determined by Amsel's (green box) and Nugent's (Score = 0) criteria. Participant F was re-colonized with A. vaginae, Leptotrichia & Sneathia spp. and G. vaginalis; BV was diagnosed at that time. The green line at the bottom of the graph shows days of metronidazole treatment and the red line denotes menses.

Figure 4. Bacterial fluctuations in a participant with recurrent BV.

Figure 4A depicts dynamic patterns of BV bacteria in Participant G across a period spanning 11 months. She was diagnosed with BV by Amsel's and Nugent's criteria at entry (Episode 1, 4B), and was responsive initially to treatment. However, she had two more episodes subsequently (4C and 4D) and each time she responded to metronidazole treatment but had a return of BV-associated bacteria and went on to develop BV.