Vaginal colonisation by probiotic lactobacilli and clinical outcome in women conventionally treated for bacterial vaginosis and yeast infection

Demographic data

A total of 10 women with BV were recruited in trial I. Out of 30 women recruited in
trial II, two women dropped out from the study. Hence, there were 9 women in group-1
(BV), 9 in group-2 (R-VVC) and 10 women in group-3 (R-VVC control group) (Additional
file 1: Figure S1). The demographic data for 38 women in trial I and trial II, who completed
the study until the six-month follow up is shown in Table 1. Among the 17 women with BV in trials I and II, 14 had a history of BV. Mean length
of symptoms of BV was 14 months with a range of 1 to 36 months. All the women with
yeast infection had a history of VVC with a mean of 42 months of having symptoms (range
12–120 months).

Table 1. Demographics and clinical characteristics of the study groups at enrollment.
^a
Each group received EcoVag® capsules containing the mix of L. gasseri DSM 14869 and L. rhamnosus DSM 14870

Colonisation with EcoVag® strains

First trial: Treatment of BV

None of the women were colonised by any lactobacilli before administration of EcoVag®
capsules. Following administration, nine out of ten women were colonised by either
of the EcoVag® Lactobacillus strains during the study (Fig. 2, Additional file 2: Figure S2; Additional file 3: Table S1,). In eight women, EcoVag® strains persisted for at least two weeks after
stopping the treatment (Additional file 2: Figure S2). In three women (no. 2, 6 and 9), either of the strains was identified
at month 4 (two months after the treatment was stopped) and persisted until month
five (three months aft-er the treatment) in two of them (no. 6 and 9). L. gasseri DSM 14869 was more frequently isolated than L. rhamnosus DSM 14870 (17 vs 9 out of 42 samples) but the difference was not significant (P?=?0.098)
(Additional file 3: Table S1).

Fig. 2. Frequency of isolation of EcoVag® and other Lactobacillus strains during the treatment of BV and VVC. The frequency of isolation (y axis) was
determined as the percentage of samples positive for EcoVag® strains or other lactobacilli
on the total number of samples for each group. Trial I: women treated for BV with
antibiotics and EcoVag® (BV EcoVag®). Trial II: women treated for BV with antibiotics
and prolonged administration of EcoVag® (BV EcoVag®), women treated for VVC with fluconazole
and prolonged administration of EcoVag® (VVC EcoVag®), women treated for VVC with
fluconazole only (VVC fluconazole). The difference in frequency of isolation of EcoVag®
strains or other lactobacilli between groups was compared by Fisher’s exact test.
*** P??0.001

Second trial: Treatment of BV and yeast infection with long-term administration of
EcoVag®

In the second trial, we evaluated if increasing the EcoVag® dose frequency would improve
colonisation with L. rhamnosus DSM 14870 and L. gasseri DSM 14869 in BV patients and if the administered EcoVag® strains could colonise yeast
infected patients, ie. when the microbiota has not been disturbed with antibiotics.
We also increased the number of samples collected from one to two between each menstruation
as the level of lactobacilli may vary during the menstruation cycle. Sample 1 was
taken after menstruation (day 7) and immediately before the first weekly EcoVag® treatment.

In group-1 with BV positive women, seven out of nine women regularly sent the samples
until month six except for woman no. 13 and 18 (Additional file 2: Figure S2). In five of these women (no. 4, 14, 15, 17 and 20) either of the EcoVag®
strains was identified until month five. Overall, L. gasseri DSM 14869 was isolated more frequently (32 out of 86 samples, 37 %) than L. rhamnosus DSM 14870 (9 out of 86 samples, 10 %) (P??0.001) (Fig. 2, Additional file 3: Table S1).

Non-EcoVag® lactobacilli were in majority identified in 63 out of 86 samples (73 %)
from 9 women compared to EcoVag® strains identified in 41 samples (48 %) (P??0.002)
(Fig. 2, Additional file 3: Table S1). Interestingly, four women (no. 2, 10, 14 and 18) diagnosed with BV had
lactobacilli at the time of recruitment and EcoVag® strains were more frequently isolated
from women that did not have lactobacilli at the start of the study (29 out of 48
samples (60 %) from 5 women) compared to women who had lactobacilli at the start of
the treatment (12 out of 38 samples (31.5 %) from four women) (P??0.01) (Additional
file 2: Figure S2). No significant difference was observed in the frequency of isolation
of EcoVag® strains or other lactobacilli between sample 1 (day 7) and sample 2 (day
21).

A comparison of both trials I and II shows that colonisation by EcoVag® lactobacilli
in women with BV did not improve significantly upon increasing the dose (Fig. 2). For this comparison, only sample I of trial II was included as it corresponds to
the sample taken in trial I. In month 4, EcoVag® strains were isolated in more women
in trial II (6 out of 8) than in trial I (3 out of 8) but this was not significant
(P?=?0.31) (Additional file 2: Fig. S2). Overall, EcoVag® strains were identified in 24 out of 42 (57 %) samples
in trial I, and 24 out of 48 (50 %) samples in trial II. The frequency of isolation
of other Lactobacillus species was significantly higher in the second trial (34 out of 48 samples, 71 %)
than in the first trial (10 out of 42 samples, 24 %) (P??0.001). The frequency of
isolation of any lactobacilli (EcoVag® strains or other Lactobacillus strains) was similar for trials I and II (33 out of 42, 79 % vs 45 out of 48, 94 %,
P?=?0.06).

In the Candida infected patients (group-2), EcoVag® strains were isolated from 8 out of 9 (89 %)
women at some time point during the study but the Lactobacillus microbiota was dominated by non EcoVag® Lactobacillus strains (Fig. 2, Additional file 2: Figure S2). Overall L. rhamnosus DSM 14870 and L. gasseri DSM 14869 were isolated in 16 (19 %) and 10 (12 %) out of 86 samples respectively
(Fig. 2, Additional file 3: Table S1). Furthermore, as observed before, EcoVag® strains were more frequently
isolated from women that did not have lactobacilli at the start of the study (15 out
of 31 samples (48 %) from 3 women) compared to women from whom lactobacilli were isolated
(9 out of 53 samples (17 %) from 6 women) (P??0.01).

A comparison of group-1 and group-2 in trial II showed that EcoVag® lactobacilli were
isolated more frequently in BV patients pretreated with antibiotics (41 out of 86
samples, 48 %) than from Candida infected patients pretreated with anti-fungal drugs (24 out of 86 samples, 28 %)
(P??0.05).

In Candida infected patients receiving fluconazole only (group-3), lactobacilli were regularly
isolated from 9 out of 10 women. Other Lactobacillus species isolated in all three groups in trial II were L. gasseri, L. crispatus, L. iners, L. jensenii, L. vaginalis, L. reuteri, paracasei, L. plantarum, L. rhamnosus, L. fermentum, L. acidophilus and L. salivarius (Additional file 3: Table S2). L. gasseri was the most frequently isolated species in women with BV receiving EcoVag® while
L. crispatus and L. gasseri were the most prevalent species in women with VVC receiving EcoVag® (group-2) or
fluconazole only (group-3). A significantly higher proportion of samples positive
for L. crispatus were found in women with C. albicans infection (group-2 and group-3) (Additional file 3: Table S3) but a higher proportion of women in group-2 and group-3 were already positive
for L. crispatus at the beginning of the study.

Clinical Outcome

First trial: Treatment of BV

Among the 10 women in this group, two women did not visit the clinic for the 6 month
follow up and were therefore excluded from the analysis on clinical outcome. The cure
rate after month 6 was 50 % with four women cured and four with a relapse of BV (Table 2, Fig. 3).

Table 2. Clinical outcome

Fig. 3. Cure rate of BV and VVC following different treatments. Two women in Trial I and three
women in Trial II group-1 had a new sexual partner during the study and developed
symptoms and experienced relapse of BV

The frequency of isolation of EcoVag® strains was slightly higher in cured women (14
out of 20 samples) than in women with relapse (6 out of 14 samples) but was not significant
(P?=?0.16) (Fig. 4, Additional file 3: Table S4). However, the frequency of isolation of any lactobacilli was associated
with the cure of BV (Fig. 4, Additional file 3: Table S4). Nineteen out of 20 (95 %) samples from the four cured women contained
either EcoVag® strains or other Lactobacillus strains compared to only 6 samples out of 14 (43 %) from women who experienced a
relapse (P??0.01).

Fig. 4. Association between the frequency of isolation of EcoVag® and other Lactobacillus strains with cure of BV and VVC (until 6-month follow up). The frequency of isolation
(y axis) is determined as the percentage of samples positive for EcoVag® strains and
other lactobacilli on the total number of samples for each group (cured or relapse).
Trial I: women treated for BV with antibiotics and EcoVag® (BV EcoVag®). Trial II:
women treated for BV with antibiotics and prolonged administration of EcoVag® (BV
EcoVag®), women treated for VVC with fluconazole and prolonged administration of EcoVag®
(VVC EcoVag®), women treated for VVC with fluconazole only (VVC fluconazole). The
difference in frequency of isolation of EcoVag® strains or other lactobacilli between
cured women and woman with relapse (until 6-month follow up) was compared by Fisher’s
exact test. *P??0.05, **P??0.01, ***P??0.001

Second trial: Treatment of BV and yeast infection

Nine women in group-1 completed the study and the cure rate for BV after month 6 and
12 was 67 % with 3 relapses and 6 cured women (Table 2, Fig. 3). An association was found between the cure of BV and both the presence of EcoVag®
strains or any lactobacilli. The frequency of isolation of EcoVag® strains was higher
in cured women (33/54 samples, 61 %) than in women with relapse (8/32 samples, 25 %)
(P??0.01) (Fig. 4, Additional file 3: Table S4). Furthermore, all the 54 samples (100 %) from cured women were identified
with lactobacilli compared to 25 out of 32 (78 %) samples in women with relapse (P??0.001).

Three women who had new sexual partner during the study developed symptoms and experienced
relapse while the others (six women) with their same partners remained cured (P??0.05).
Overall in both trials, among 17 women treated for BV, 5 women changed their partners
during the follow up and experienced relapse while only one woman that did not change
partner had a relapse. Thus, the change of sexual partner was associated with the
relapse of BV, OR 77 (2.665 to 2225 at 95 % CI, P?=?0.0014).

All nine women infected with Candida and treated with fluconazole and EcoVag® (group-2) were cured at the 6-month follow
up (Table 2, Fig. 3). All the treated women in this group visited the clinic for a second follow up (12–18
months after the initial treatment) and eight were still cured (89 % cure rate).

In the group receiving fluconazole only (group-3), all the women were also cured at
the 6-month follow up but three women got a relapse before 12-months follow up and
were given a new course of fluconazole treatment plus EcoVag® (Table 2, Fig. 3). The 12-month cure rate was thus slightly higher in women receiving fluconazole
and EcoVag® (89 %) than in women receiving fluconazole only (70 %) but this was not
statistically significant (P?=?0.582). No association was found between cure of VVC
and isolation of EcoVag® strains or other lactobacilli.

Adverse events

Only a few adverse events were reported during the intervention period in both trials.
In trial I, one man treated with clindamycin had a skin rash and stopped taking the
antibiotic after six days. In trial II, women 6 and 9 in group-2 stopped taking the
last 10 and 2 capsules respectively.