Characteristics associated with Lactobacillus-depleted vaginal microbiota in women with different sexual behavior

Abstract

The aim of this study was to compare the vaginal microbiome of women with different sexual behaviors and to examine characteristics associated with Lactobacillus-deprived community state type IV. In this prospective study performed in a community-based population in a city of approximately 150,000 inhabitants in Southeast region of Brazil, vaginal swabs were obtained of 109 participants, including women who only had sex with women (n = 54) and women who only had sex with man (n = 55). Sociodemographic data, sexual and intimated hygiene practices of the participants were also assessed. Vaginal microbiota was assessed by sequencing the hypervariable regions V3 and V4 of 16 S ribosomal nucleic acid gene (Illumina 250 PE). Alpha diversity (Shannon index) was compared between the two groups by the Mann-Whitney test. Logistic regression analyses were performed to estimate odds ratios (OR) and 95% confidence intervals (CI) for the association between covariates with the Lactobacillus-deprived community state type IV. Results showed that overall distribution of vaginal CSTs did not differ between the two study groups (P = 0.19). However, alpha diversity was increased in women that only have sex with women (P = 0.0018). Lactobacillus-deprived community state type IV was associated with lower income (OR: 4.15, 95% CI: 1.04–16.46) and use of sex toys (OR: 3.97, 95% CI: 1.09–14.45). In conclusion, nearly one-third of women that only have sex with women had a sub-optimal vaginal microbiome and show evidence of sex transmissibility of CST-IV associated organisms.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-34977-2.

Introduction

Predominance of Lactobacillus sp. in a low-diversity environment remains as the most accepted definition of a healthy state of the vaginal microbiota¹. Loss of vaginal lactic acid-producing Lactobacillus is accompanied by increased diversity reflecting the overgrowth of strict and facultative anaerobes, as in bacterial vaginosis². Women with bacterial vaginosis are at increased risk for several sexually transmitted infections (STI), including, but not restricted to, human papillomavirus (HPV) and human immunodeficiency virus (HIV)^3,4. The symptoms often associated with bacterial vaginosis, such as unpleasant odor and increased vaginal discharge, also have a substantial impact on women’s physical and psychological well-being⁵.

Occurrence of bacterial vaginosis is associated with several sociodemographic and behavior characteristics of the population^6,7. Some sexual practices that are frequently reported by women who have sex with women (WSW) were already associated with bacterial vaginosis, such as receptive oral sex and utilization of sex toys^8–11. In fact, some studies have suggested the sexual transmissibility of bacterial vaginosis in WSW^8,12. This is particularly problematic given the high rates of STI observed in WSW and the scarce number of studies performed with this population^13,14.

Sequencing of 16s ribosomal ribonucleic acid (rRNA) gene allowed a better understanding of the composition of the vaginal was achieved in the past decade. These molecular-based studies have showed that the vaginal microbiota of nearly all women of reproductive age fit into one of the five community state types (CSTs I to V) already described^2,15. While CSTs I, II, III and V are dominated by, respectively, Lactobacillus crispatus, L. gasseri, L. iners and L. jensenii, CST IV is Lactobacillus-deprived and comprise most of the bacterial vaginosis cases². Literature suggest that the four Lactobacillus-dominated CST are not equally protective for the vaginal environment, as the protective role of L. iners has been largely questioned^16,17. In Brazil, nearly two-thirds of women of reproductive age presents CST III or CST IV¹⁵. This finding leads to several concerns about the impact that the vaginal microbiota might have in this population, particularly in terms of increased rates of STI transmission.

Despite the numerous studies utilizing molecular tools to assess the vaginal microbiota that have been performed in the past decade, a very limited number focused in the population of WSW^18–22. An even lower number of studies assessed women reporting exclusive female partnerships^21,22. Given the importance of better understanding the vaginal microbiota of WSW and the contributing factors for loss of Lactobacillus-dominance, this study aimed to molecularly compare the vaginal microbiota of WSW with those women who only had sex with men (WSM) in the 12 months prior to their enrollment. Additionally, the association of sociodemographic, sexual practices, and clinical characteristics with the Lactobacillus-depleted CST IV was investigated.

Results

Sociodemographic data, sexual and intimated hygiene practices

Sociodemographic characteristics of the 109 participants are displayed in Table 1, according to the study group. Most of the participants (n = 97, 90.0%) had less than 40 years of age at enrollment. The group of WSW had a greater proportion of participants with higher educational level (at least high school degree) (P < 0.0001). Regarding the marital/partnership status, more than half of WSM (n = 37, 67.3%) were in a regular partnership, differing of WSW group with 63% (n = 34) of participants that were not in a regular partnership (P = 0.002). Smoking cigarettes was reported more frequently by WSW (P = 0.001). Study groups did not differ regarding the percentage of women with formal/informal jobs and with lower family incomes (P > 0.05). All WSM included in the study were covered by exclusively the Brazilian public healthcare system (P < 0.0001), differing from the WSW, among which 68.5% (n = 37) also had access to private health services/clinics.

Table 1.

Sociodemographic characteristic of the population, according to the study groups.

Characteristic	WSM (n = 55)	WSW (n = 54)	P-valuea
Age (years)			0.23
< 40	47 (85.5)	50 (92.6)
40+	8 (14.5)	4 (7.4)
Skin color			0.89
White	36 (65.5)	36 (66.7)
Black/other	19 (34.5)	18 (33.3)
School (y)			< 0.0001
< 11	20 (36.4)	2 (3.7)
11+	35 (63.6)	52 (96.3)
Marital/partnership status			0.002
Regular	37 (67.3)	20 (37.0)
Non-regular	18 (32.7)	34 (63.0)
Smoking cigarettes			0.001
No	46 (83.6)	29 (53.7)
Yes	9 (16.4)	25 (46.3)
Have formal/informal job			0.48
No	13 (23.6)	16 (29.6)
Yes	42 (76.4)	38 (70.4)
Low family income b			0.80
No	48 (87.3)	48 (88.9)
Yes	7 (12.7)	6 (11.1)

WSM: Women who have sex with men; WSW: women who have sex with women.

^aChi-squared or Fisher exact test.

^bIncome inferior to one half of the minimum wage by family member.

The two study groups were also compared by sexual and intimate hygiene habits, as displayed in Table 2. As expected, the sexual practices reported by participants differed between the groups. Utilizing sex toys during sex was only reported by participants of the WSW group (n = 19, 35.2%) (P < 0.001). All WSM participants reported to practice receptive penile-vaginal sex (P = 0.01), while receptive oral sex was reported by nearly all participants of WSW (P < 0.001). Participants reporting to have condomless sex (none or occasional condom use) did not differ between the groups (P = 0.10). Use of hormonal contraceptives was only reported by WSM (P < 0.0001).

Table 2.

Sex practices, intimate hygiene habits and type of contraceptive methods, according to the study groups.

Characteristic	WSM (n = 55)	WSW (n = 54)	p-valuea
New recent sex partner (< 3 months prior to enrollment)			0.75
No	48 (87.3)	48 (88.9)
Yes	7 (12.7)	6 (11.1)
Use of sex toys			< 0.001
No	55 (100.0)	35 (64.8)
Yes	0 (0.0)	19 (35.2)
Has receptive penile-vaginal sex			0.01
No	0 (0.0)	6 (11.1)
Yes	55 (100.0)	48 (88.9)
Has receptive anal sex (penile, digital, toys)			0.03
No	36 (65.5)	45 (83.3)
Yes	19 (34.5)	9 (16.7)
Has receptive oral sex			< 0.001
No	25 (45.5)	1 (1.8)
Yes	30 (54.5)	53 (98.2)
Intimate douching			0.11
No	47 (85.2)	39 (72.7)
Yes	8 (72.7)	15 (27.3)
None/occasional condom use)			0.10
No	41 (74.6)	47 (87.0)
Yes	14 (25.5)	7 (13.0)
Use of hormonal contraceptives			< 0.0001
None	20 (36.4)	54 (100.0)
Oral	27 (49.1)	0 (0.0)
Inject	8 (14.5)	0 (0.0)

WSM: Women who have sex with men; WSW: women who have sex with women.

^aChi-squared or Fisher exact test.

Prevalence of cervicovaginal infections

Results in Table 3 show that prevalence of infections of the lower genital tract caused by any HPV genotypes, high-risk HPV genotypes, Chlamydia trachomatis, Trichomonas vaginalis and Candida spp. did not differ between WSM and WSW. Infection by T. vaginalis was only reported in the group of WSW (n = 3, 5.6%). Status of vaginal microbiota assessed by microscopy differed between the groups (P = 0.01). Significantly increased number of bacterial vaginosis cases detected by microscopic was observed among WSW (n = 21, 39.6%) when compared to WSM (n = 8, 14.8%) (P = 0.004).

Table 3.

Infections of the lower genital tract and status of vaginal microbiota, according to the study groups.

Characteristic	WSM (n = 55)	WSW (n = 54)	p-valuea
HPV (any genotype) b			0.20
No	20 (37.7)	27 (50.0)
Yes	33 (62.3)	27 (50.0)
High-risk HPV b, c			0.13
No	33 (62.3)	41 (75.9)
Yes	20 (37.7)	13 (24.1)
Chlamydia trachomatis			0.57
No	53 (96.4)	53 (98.2)
Yes	2 (3.6)	1 (1.8)
Trichomonas vaginalis			0.08
No	55 (100.0)	51 (94.4)
Yes	0 (0.0)	3 (5.6)
Candida spp. (morphotypes detected at microscopy)			0.28
No	49 (89.1)	51 (94.5)
Yes	6 (10.9)	3 (5.5)
Vaginal microbiota status (Nugent´s scores) c, d			0.01
Normal (0–3)	40 (74.1)	27 (50.9)
Intermediate (4–6)	6 (11.1)	5 (9.4)
Bacterial vaginosis (7–10)	8 (14.8)	21 (39.6)

^aChi-squared or Fisher exact test.

^b Data missing from two WSM participants.

^cHPV16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, 66, and/or 68); ^cData missing from two participants (one at each study group).

^dBased on semi-quantitation of bacterial morphotypes on Gram-stained smears at 1000x magnification. WSM: Women who have sex with men; WSW: women who have sex with women.

Vaginal microbiome and associations

Results of the molecular characterization of the vaginal microbiota are displayed in Fig. 1; Table 4. The L. iners-dominated CST III was the most frequent in this population (37.6%, n = 41) and were similarly distributed in both study groups (Table 4). Second most frequent CST was CST I (dominated by L. crispatus) detected in 37 (33.9%) participants. The other Lactobacillus-dominated CSTs, CST II and CST V, were under represented in the study population (5.6%, n = 6). Presence of the Lactobacillus-deprived CST IV was observed in 25 (22.9%) participants. Overall distribution of CSTs did not differ between WSM and WSW (P = 0.19). Frequency of Lactobacillus-depleted CST IV in the group of WSW (29.6%, n = 16) was higher, but not statistically different, than WSM (16.4%, n = 9) (P = 0.10, at chi-squared test).

Fig. 1.

Heatmap of log10-transformed relative abundance of those taxa with a minimum of 5% overall prevalence in the population. Superior bars represent (i) community state types (CST) I to V assigned using³⁵; (ii) groups of women who have sex with women (WSW) and women who have sex with men (WSM) and; (iii) results of microscopic classification according to Nugent´s scores³². Color legend for the taxa relative abundance, CSTs, study groups and Nugent´s scores are indicated in the right side of the figure.

Table 4.

Distribution of vaginal community state types, according to the study groups.

Community state types (CSTs)	Overall (n = 109)	WSM (n = 55)	WSW (n = 54)	p-valuea
CST I (Lactobacillus crispatus dominance)	37 (34.0)	22 (40.0)	15 (27.8)	0.192
CST II (L. gasseri dominance)	1 (0.9)	0 (0.0)	1 (1.8)
CST III (L. iners dominance)	41 (37.6)	20 (36.4)	21 (39.0)
CST IV (Lactobacillus depleted)	25 (22.9)	9 (16.3)	16 (29.6)
CST V (L. jensenii dominance)	5 (4.6)	4 (7.3)	1 (1.8)

^aChi-squared test. WSM: Women who have sex with men; WSW: women who have sex with women.

However, as displayed in Fig. 2, WSW showed significantly increased alpha diversity of the vaginal microbiome (Median = 1.322, Min-Max: 0.045–2.412), measured utilizing the Shannon index, when compared to WSM (Median = 1.048, Min-Max: 0.056–2.391) (P = 0.0018).

Fig. 2.

Alpha-diversity measured using Shannon index between the two study groups. P-value at Mann-Whitney non-parametric test.

The association between the overall population variables and Lactobacillus-depleted CST IV is shown in Table 5. Univariate analysis showed that CST IV was associated with low family incomes (OR: 3.47, 95% CI: 1.05–11.53) and use of sex toys during intercourse (OR: 4.16, 95% CI:1.45–11.90). Both associations were confirmed after performing a multivariable analysis. An additional univariate analysis was performed exclusively with WSW (Supplementary Table 1) and showed that use of sex toys (OR: 3.60, 95% CI:1.06–12.24) was associated with CST IV in this group.

Table 5.

Odds ratios (OR) and 95% confidence intervals (95% CI) for the association between sociodemographic, behavioral, and clinical characteristics of the study population with Lactobacillus-depleted CST IV vaginal microbiota.

Characteristic	Crude		Multivariable**
	OR (95% CI)	p-value	OR (95% CI)	p-value
Study group		0.104		0.795
WSM	1.00		1.00
WSW	2.15 (0.85–5.41)		0.85 (0.26–2.78)
Age (y)		0.436
< 40	1.00
40+	0.54 (0.11–2.57)
Marital status		0.164
Regular partnership/married	1.00
Single	1.91 (0.77–4.73)
Ethnicity		0.805
White	1.00
Black/other	1.13 (0.44–2.86)
School (y)		0.051		0.081
< 11	1.00		1.00
11+	7.75 (0.99–60.83)		6.91 (0.79–60.49)
Low family income*		0.042		0.043
No	1.00		1.00
Yes	3.47 (1.05–11.53)		4.15 (1.04–16.46)
New sex partner (< 3 months)		0.495
No	1.00
Yes	0.57 (0.12–2.80)
Smoking cigarettes		0.921
No	1.00
Yes	1.05 (0.40–2.74)
Hormonal contraceptive		0.648
None	1.00	0.827
Oral	1.27 (0.46–3.53)
Injectable	1.21 (0.22–6.57)
None/occasional condom use		0.496
No	1.00
Yes	0.69 (0.23–2.02)
Douching (intimate)		0.686
No	1.00
Yes	1.24 (0.43–3.59)
Use of sex toys		0.008		0.036
No	1.00		1.00
Yes	4.16 (1.45–11.90)		3.97 (1.09–14.45)
Receptive penile-vaginal sex		0.709
No	1.00
Yes	1.52 (0.17–13.6)
Receptive anal sex (penile, digital, toys)		0.214
No	1.00
Yes	0.47 (0.15–1.53)
Receptive oral sex		0.299
No	1.00
Yes	1.86 (0.57–6.03)
High-risk HPV infection		0.726
No	1.00
Yes	0.84 (0.31–2.25)
Chlamydia trachomatis and/or Trichomonas vaginalis infection		0.119		0.220
No	1.00		1.00
Yes	3.77 (0.71-20.00)		3.47 (0.47–25.39)

*Incomes inferior to one half of the minimum wage by family member.

**Variables with P-value < 0.15 included in the multivariable analysis. WSM: Women who have sex with men; WSW: women who have sex with women.

Relative abundances of Gardnerella, Prevotella and Fannyhessea were compared between WSM and WSW (Fig. 3A). Statistically significant higher relative abundance of Prevotella and Fannyhessea were observed in vaginal samples from WSW (P < 0.05). The 75th percentile value of the relative abundance in WSM for three genera were then defined as the cut-off valued for ´high´ taxa abundance. As observed in Fig. 3B, WSW significatively were more likely to have high relative abundance of Gardnerella (OR: 2.72, 95% CI: 1.21–6.10), Prevotella (OR: 3.95, 95% CI: 1.75–8.89) and Fannyhessea (OR: 3.15, 95% CI: 1.41–7.08).

Fig. 3.

Comparison of the three taxa with main roles in the initiation of the pathogenesis of bacterial vaginosis between women who have sex with men (WSM) and women who only have sex with women (WSW). (A) Relative abundances of the taxa were compared between the groups utilizing Mann-Whitney non-parametric test. Purple horizontal bars indicate median values and blue bars represent the 25th (inferior) and 75th (superior) percentiles. (B) Odds ratios (OR) and 95% confidence intervals (95% CI) for the association between taxa high relative abundances and WSW. Cut-off values for high relative abundances were defined by the 75th percentile taxa in the group of MSM: Gardnerella spp. = 1.14%, Prevotella spp. = 0.33%, and Fannyhessea = 0.05%. ^*Only Fannyhessea vaginae was considered.

Discussion

To the best of our knowledge, this is the first full characterization of the vaginal microbiota utilizing 16 S rRNA sequencing of a population of WSW with exclusive female sexual partners. Also, this study shows that some population sociodemographic and sexual characteristics are independently associated with Lactobacillus-deprived microbiota (i.e. CST IV). Despite the small sample size (n = 109), the inclusion criteria adopted for this study were very strict to assure that all participants allocated at WSW group only had female sexual partners in the 12 months prior to enrollment, while WSM had only male partners.

Participants of the two study groups were homogeneous for characteristics as age, ethnicity, family income among others. However, WSW showed increased education levels when compared to WSM. This was not the same observed in other populations in which WSW showed same or inferior levels when compared to women that have male partners^9,23. The rate of participants living in non-regular relationships (single, not living with partner) was increased among WSW, which is in agreement with other study by Olson et al. (2018) that showed a higher frequence of never married participants among WSW when compared to those that had/have male partnerships⁹. Also, the use of tobacco (cigarette smoking) was higher in WSW, showing with similar rates (~ 45%) reported by the literature in MSM from other countries^10,24. As expected, some sexual practices were differently reported by the two study groups, as receptive oral sex, sex toys, condomless sex. These findings were expected based on several previous reports of sexual practices among women with female sexual partners^8,13,21,25. Although the prevalence rates of STIs in this study did not differ between the groups, presence of Nugent bacterial vaginosis was significantly increased in WSW. In fact, there is a consensus that bacterial vaginosis is increased among WSW^9,13,18,26.

While the presence of microscopic bacterial vaginosis is well documented among WSW and should draw more attention to the study of the microbiota of this population, there are few studies available that utilized molecular methods for this aim^18–22. Of these, only two were based on 16 S rRNA sequencing that has been the most accepted molecular tool for the assessment of the vaginal microbiota^20,21. By sequencing regions of 16 S rRNA, vaginal microbiota may be clustered into CSTs allowing the differentiation between Lactobacillus-dominated and -deprived microbiota. This study showed a higher prevalence of CST IV in WSW when compared to WSM, although not significantly. A study by Plummer et al. (2019) utilized CST clustering for assessment of WSW and showed that Lactobacillus-deprived CST microbiota is highly prevalent and incident among WSW²¹. However, it is difficult to compare their results with the data shown in this study as (i) their study design did not comprise a group of WSM and (ii) several WSW had reported recent sex intercourse with men. On the other hand, the same study by Plummer et al. (2019) also evaluated WSW longitudinally and showed that this population present a high temporal variability of vaginal CSTs, as previously demonstrated in WSM by Gajer et al. (2012)^16,21. Additionally, the two studies are consonant in showing that L. iners-dominated CST III is more unstable when compared to other Lactobacillus-dominated CSTs^16,21. In addition to the temporal instability of vaginal microbiota, CST III is also associated with increased susceptibility to STIs¹⁷. This is particularly important, since it is the most frequent CST observed in some populations, including the Brazilian WSM and WSW²⁷.

In relation to the second aim of this study, the association of sociodemographic, behavioral, and clinical covariates of this population with Lactobacillus-depleted CST IV was examined. This study failed to demonstrate some associations that were previously observed in WSW with bacterial vaginosis, as ethnicity, smoking cigarettes, new recent female partners, among others^10,14. It should be considered that the restricted sample size of this study may have contributed negatively for this analysis, but we opted not to include women with male and female sexual partnerships. Having an individual income equal or inferior to half of the minimum wage at the time of enrollment income was independently associated with CST IV. The association between household income and CST IV was not sufficiently investigated, but a recent study conducted with more than 500 participants failed to demonstrate this association²⁸. Of noteworthy, the type of sexual partnerships was not considered for inclusion criteria and it was not addressed at the latter study, limiting the comparison with the current study.

Importantly, the study by Plummer et al. (2019) showed that sharing sex toys by WSW was associated with CST IV with CST I as reference²¹. The results of the current study on the independent association between use of sex toys and presence of CST IV corroborate their previous findings. According to evidences on the literature, exchange of bodily fluids containing biofilms formed by bacterial vaginosis-associated bacteria is the most likely explanation for this association^7,21,29. Indeed, the current data reinforce this hypothesis given that WSW are more likely to have high vaginal abundances of the three genera that are most accepted as the most important for the pathogenesis of bacterial vaginosis³⁰. Collectively, the results of this study point out to a high prevalence of a sub-optimal microbiota (CST III and CST IV) in WSW and add important evidence of the sexual transmissibility of dysbiosis-associated species. These observations may contribute for improving the clinical management of bacterial vaginosis in WSW and their female partners.

Methods

Participants selection, characteristics, and sampling procedures

Institutional Review Board of the Botucatu Medical School reviewed and approved the study (CAAE: 81601024.9.0000.5411 - nº 7.177.825). The study goal and procedures were explained to all participants that, upon agreement, signed a written consent form. Participants were informed that they could drop the study at any stage. This study adhered to the ethical principles outlined in the Declaration of Helsinki of Ethical Principles for Medical Research Involving Human Participants.

Cis-gender women of at least 18 years-old, in reproductive age and with active sex life were considered for the study enrollment. All participants were enrolled in Botucatu city (of approximately 150,000 inhabitants) in São Paulo state, Brazil. A total of 109 women were included in the study and comprised the two study groups: WSW (n = 54) and WSM (n = 55) in the 12 months prior to the study. All participants were enrolled in Botucatu, São Paulo, Brazil, from 2015 to 2020. Women attending community-based clinics for routine Pap-testing were considered for study enrollment in the WSM group. The participants of the WSW group were included in one primary health care unit after seeking this clinic for participating of the study. They volunteered to participate of the study, as it was publicized in different forms: posts in social media, posters and flyers distributed at commercial areas, LGBT activism groups, and feminist collectives.

Prior to sampling, participants answered to face-to-face interview using a structured form. All interviews were performed by one member of the study group in a closed room for privacy. After that, physical examination was performed. Using a non-lubricated speculum, two swabs were taken from the mid-vaginal wall. First swab (Copan, Brescia, Italy) was immediately placed into liquid Amies transport media and stored at -80° C for molecular assessment of the vaginal microbiota. The second swab was utilized for preparing vaginal smears for further microscopic analysis Gram-stained. Cervical samples were obtained for Chlamydia trachomatis molecular testing. Contents of the posterior fornix was taken using – plastic spatula and immediately inoculated in modified Diamond´s culture media for Trichomonas vaginalis.

Detection of lower genital tract infections

Vaginal smears were Gram-stained and screened for the presence of Candida spp.-like morphotypes. When in the presence of yeast budding and/or pseudohyphae and/or true hyphae, participant was considered positive for Candida infection. Total DNA extraction of cervical samples was performed for C. trachomatis molecular testing by taxon-directed polymerase chain reaction, as described elsewhere³¹. Culture for T. vaginalis was performed in modified Diamond’s medium at 37 °C in a 5% CO₂ environment. Cultures were evaluated daily by fresh mount microscopy and, if negative, re-incubated for up to 72 h. Only at the presence of motile trophozoites at microscopy samples were considered as positive. Negative cultures after 72 h were discarded and were considered negative for T. vaginalis.

Assessment of vaginal microbiota

Microscopy

Gram-stained vaginal smears utilized for microscopic detection of Candida sp. morphotypes were also utilized for classification of the vaginal microbiota according to Nugent´s scoring system³². Samples were observed at 1000x magnification and categorized as normal microbiota (scores 0 to 3), intermediate microbiota (scores 4 to 6) or bacterial vaginosis (score 7 to 10).

Sequencing of V3V4 regions of 16 S rRNA gene

Vaginal swabs inserted into Amies media (Copan, Brescia, Italy) were allowed to melt at room temperature, vortexed and pressed against the tube walls to detachment of any content adhered to it. Extraction of total bacterial DNA was performed utilizing ZymoBiomics DNA Kit (Zymo, Irvine, CA) according to manufacture´s instructions. Amplification of V3V4 regions of 16s rRNA genes was performed utilizing the 357 F and 805R set of primers (synthetized on demand and HPLC purified by Sigma) and GoTaqG2 kit (Promega, Madison, WI). Amplicons were then quantified and adjusted to 2nM (Quantus dsDNA kit, Promega, Madison, WI) and further purified (PureLink Quick PCR Purification, Invitrogen, Waltham, MA). Sequencing reactions were prepared utilizing o MiSeq Reagent 500V2 (Illumina), according to manufacture´s instructions to Illumina MiSeq platform 250 PE with expected sequences overlap of ~ 50 bp.

A total of 218 .fastq files were generated containing raw forward and reverse sequences. Files were then imported to software R and analysed using ´dada2´³³. Sequences were quality filtered (truncQ = 2) and dereplicated. A parametric error model was applied and errors were removed. Sequences were checked for chimera. Taxonomic assignments were performed at species level utilizing Silva database (version 138) with addition of sequences from BVAB1 and BVAB2. Amplicon sequence variants (ASVs) with sum resulting in less than 100 reads among all study samples were excluded from the analysis. Relative abundances were calculated by dividing the number of reads of a certain ASV by the total reads in that sample and further multiplied by 100. Alpha-diversity was calculated utilizing Shannon index in the ´vegan´ package for R³⁴. Samples were clustered into CSTs utilizing the VALENCIA (VAginaL community state typE Nearest CentroId clAssifier), a tool that categorizes samples by their resemblance to a predefined set of reference centroids divided in 13 CSTs (CST I-A and I-B, CST II, CST III-A and III-B, CST IV-A, IV-B and IV-C and CST V)³⁵.

Statistics

The frequency of sociodemographic, behavioral, and clinical characteristics among WSW and WSM participants were calculated and compared using Chi-squared test or Fisher exact test, with P-values < 0.05 considered to be statistically significant. Univariate logistic regression analysis was utilized to estimate odds ratios (OR) and 95% confidence intervals (CI) for the association between population covariates with the presence of a Lactobacillus-depleted microbiota (i.e. CST IV) in relation to the Lactobacillus-dominated CSTs. Those variables with P-value < 0.15 at univariate analysis were then added to a multivariable logistic regression model with CST IV as the independent variable. These analyses were performed using Stata/SE version 17 (StataCorp, College Station, TX). Alpha-diversity and relative abundances of Gardnerella, Prevotella and Fannyhessea were compared by study group utilizing the non-parametric Mann-Whitney test (P-values < 0.05 as statistically significant) utilizing GraphPad Prism version 9 (GraphPad, La Jolla, CA).

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

MAOI: participants enrollment, funding, IRBs clearance, writing the first version of the manuscript; CM: contextualization, statistical analysis, critical analysis of the results, assistance in writing the final version of the manuscript; LTB: bioinformatic analysis, critical analysis of the results, revision of the final version of the manuscript; EB: sequencing analysis, critical analysis of the results, revision of the final version of the manuscript; RB: bioinformatic analysis, data representation, revision of the final version of the manuscript; MGS: study design, critical analysis of the results, revision of the final version of the manuscript; MTCD: conception and design, funding, IRBs clearance, revision of the final version of the manuscript.

Funding

This work was supported by Sao Paulo Research Foundation (FAPESP), grants: 2015/04224–6 e 2018/19649–0. MAOI is supported with a scholarship from Sao Paulo Research Foundation (FAPESP) Grant: 2018/14770–6.

Data availability

The raw sequencing data are publicly available in the NCBI Sequence Read Archive (SRA) under the Bioproject number SAMN19760319.

Declarations

Competing interests

The authors declare no competing interests.

Details of ethical approval

Institutional Review Board of the Botucatu Medical School reviewed and approved the study (CAAE: 81601024.9.0000.5411 - no. 7.177.825). All participants were informed about the study aims and procedures and signed a consent form.