Table 1.
Classification systems of Gardnerella isolates.
Table 2.
Characteristics of BV patients and outcome groups at enrollment.
Fig 1.
LbRC5 scores vary between clinical outcome groups.
A. Mean LbRC5 scores among patients of refractory, recurrent, and remission outcome groups. Red samples = refractory patients. Blue samples = recurrent patients Green samples = remission patients. Data points are means of the log10 percent titers (= geometric means of the percents) with standard deviations. The P value was determined by mixed modeling for time dependent increases in remission versus recurrent scores. B. Fraction of patients in each clinical outcome group achieving LbRC5 scores ≥ 2 per day. Linear regression analysis (lines) indicated reasonable fits to remission and recurrent data with a common slope of 0.035, significantly different than refractory data (p = 0.002). Refractory data fit poorly to its regression line, indicating scatter (slope not significantly different from zero). C. Days from treatment initiation to LbRC5 scores >2. C. Days from initiation of treatment to sustained LbRC5 scores > 2. The first days at which LbRC5 scores were >2 for at least 2 consecutive days were recorded as hits in this Kaplan-Meier analysis; if a patient did not achieve this, she was scored as indeterminate at day 15, the day after the interval being studied. Remission scores were significantly different (p = 0.044) by log-rank analysis.
Fig 2.
Association of abundance of Gsp07 and G. swidsinskii/G. leopoldii with poor clinical response to metronidazole.
Abundance of all species of Gardnerella were determined at enrollment (1, open symbols) and as mean abundance of days 7–14 (M, closed symbols), per patient of clades by clinical outcome group. Ref = Refractory patients, Rec = Recurrent patients, Rem = Remission patients. Means and 95% CI’s are depicted. Comparing like days between outcome groups by ANOVA post-hoc Holm-Sidak, showed significant differences among Gsp07 abundance (only significant P values are shown); the arrow for G. swidsinskii/G. leopoldii comparisons indicates a significant downward trend. Paired t-tests for all 15 day 1 versus mean abundance of all clades showed no significant time-dependent changes.
Fig 3.
Number of Gardnerella species per patient.
Each species was counted if present at a titer above 10x the limit of detection, then averaged for each patient over all days after initiation of oral metronidazole therapy. The p value was determined by Kruskal-Wallis test corrected for multiple comparisons.
Table 3.
Prognosis of clinical outcomes by relative abundance of Gsp07 or G. swidsinskii/G. leopoldii, and LbRC5 scores at intervals from initiation of metronidazole therapy.