Abstract
Introduction
Global incidence of bacterial vaginosis (BV) varies widely among different population groups. In India, the prevalence in non-pregnant pre-menopausal women has been reported from 19 to 44.8%.1–3 These variations arise based on detection methods employed and differences in the socio-cultural and behavioural attributes. BV has been strongly implicated in increased risk of sexually transmitted infections (STIs), pelvic inflammatory disease (PID), preterm birth and spontaneous abortion.4–7 Most women in their child-bearing years experience symptoms of vaginitis, and its commonest clinical manifestation is BV. 1 Several variables have been evaluated as risk factors for BV, however, with conflicting results. Behaviour related to sexual activity including unprotected sex, early onset of first sexual exposure, number of lifetime sexual partners and high-risk sexual behaviour such as in sex workers are significant risk factors.8–10 Hygiene practices such as douching and menstrual protection pads also contribute to vaginal dysbiosis by depleting resident lactobacilli.11–13 While the use of intra-uterine devices (IUDs) is associated with increased risk of BV, condoms and oral contraceptive pills have been shown to have a protective effect.14,15 Identification of potential risk factors specific for different populations would help create strategies to reduce the burden of BV and its associated adverse sexual and reproductive outcomes. In view of this, the present cross-sectional study was conducted in a cohort of 283 women from Delhi in India’s northern region to identify variables having significant association with BV.
Materials and methods
Subject recruitment and sample collection
This is a cross-sectional study carried out in a tertiary care centre in Delhi, India. Between August 2018 and February 2021, 301 women (18–45 years) attending the Sexually Transmitted Diseases Clinic of the Dermatology and Venereology out-patient department (OPD) and the Obstetrics and Gynaecology OPD were seen by the first author (AC) and screened for symptoms of vaginitis. The study was approved by Institute Ethics Committee. Women who were pregnant, lactating or menstruating at the time of sample collection or received antibiotics in the preceding 4 weeks were excluded from the study. A thorough clinical examination was performed, and vaginal swabs were collected from the posterior fornix of each woman consenting to participate. Evaluation for BV was done using Nugent scoring of Gram-stained vaginal smears and referred to as Nugent-BV; a terminology proposed recently by McKinnon et al. 16 Detailed information regarding socio-demographic characteristics, clinical findings, marital, obstetric, and menstrual history and sexual behaviours were obtained in the form of pre-designed questionnaires.
Data analysis
Descriptive analysis was used to assess all variables of the study, and data were described as frequency and percentages. Continuous variables were evaluated using either Student’s t-test or Wilcoxon rank-sum test while Chi-square test or Fisher-exact were applied for categorical features. Univariate and step-wise multivariate logistic regression with probability of entry (.05) and probability of removal (.1) was performed to identify independently associated risk factors. Adjusted odds ratio and p-value were calculated. A p-value less than .05 was considered statistically significant. All statistical tests were performed using STATA 12 (StataCorp LLC).
Results
Of the 301 slides, 293 were evaluated and 8 were excluded either due to inadequate sample or reduced flora on Gram-stained vaginal smears. Excluding another 10 cases where clinical history was unavailable, 283 cases were finally included in the study. Sixty-nine (24.4%) participants were positive for Nugent-BV (Nugent score ≥7), 55 (19.4%) were intermediate (Nugent score 4–6) and 159 (65.2%) were negative for Nugent-BV (Nugent score ≤3). The overall mean age of our cohort was 32.1 ± 7.3 years. The highest proportion of positive cases was among women in the age group of 21–30 years (50.7%) and in women married for >10 years (43.3%). Higher frequency (32.8%) for Nugent-BV was noted among women with infertility (p = .02), although its duration (p = 1.00) or type (primary or secondary; p = .35) did not influence the disease status (Table 1). Other demographic factors such as duration of marriage (p = .10), gravidity (p = .72) or menstrual status (p = .46) had no significant association with BV (Table 1). Previous antibiotic use, as seen in 67 (23.7%) women, also had no significant correlation (p = .64) with Nugent-BV (Table 1). Among the study participants, 278 (98.2%) women reported having at-least one lifetime sexual partner and, 274 (96.8%) participants were sexually active in the three weeks preceding their visit (Table 2) of whom, 66 (24.1%) were diagnosed positive for Nugent-BV. Although the highest number of positive cases (42.4%) were among women who reported increased frequency of sexual intercourse (average 16–20 sexual intercourse episodes/month), the association was not statistically significant (p = .39) (Table 2). A total of 132 (46.6%) women, accounting for less than half of the study participants reported contraception including tubal ligation (19.4%), condom (17.3%), intra-uterine device (IUD) (6.0%), oral-contraceptives (3.5%), vasectomy (0.4%) and injectable-contraceptives (0.4%) (Table 2). The use of condoms had a significant negative association (p = .03) with Nugent-BV (Table 2). Interestingly, 80 (28.3%) women reported sexual intercourse in 72 h preceding their hospital visit, of whom 25 (36.8%) who had had unprotected sexual exposure tested positive for Nugent-BV (p = .01) (Table 2). The use of disposable menstrual protection products was reported by 244 (89.4%) women, although the choice (reusable or disposable) of products did not have any significant (p = .07) association with a positive diagnosis. The use of hygiene products also had no association (p = .30) with Nugent-BV. Most women screened were symptomatic (77.4%), and the presence of vaginal discharge (72.2%), lower abdomen pain (59.8%) and vaginal itching (48.9%) were the most commonly reported complaints (Table 3). However, none of the symptoms were significantly associated with Nugent-BV. Variables with a p-value < .1 (history of infertility, condom, recent unprotected intercourse and choice of menstrual protection products) were considered for logistic regression analysis. Univariate logistic regression analysis revealed higher odds for Nugent-BV among women with infertility (odds ratio (OR) = 2.02; 95% CI = 1.09–3.73, p = .02) and in women who reported unprotected sexual exposure in 72 h preceding their visit (OR = 1.00; 95% CI = 0.35–0.93, p = .02) while the use of condoms (OR = 0.38; 95% CI = 0.15–0.93, p = .03) or disposable sanitary pads (OR = 0.45; 95% CI = 0.20–0.98, p = .05) had a significant negative association with Nugent-BV (Table 4). However, multivariate regression analysis for the select variables did not identify any correlates that were independently associated with Nugent-BV.
Demographic correlates of BV among women in Delhi, India.
1Column percentage Bold values are statistically significant at α = 0.05..
Behavioural correlates of BV among women in Delhi, India.
1column percentage.Bold values are statistically significant at α = 0.05.
Clinical correlates of BV among women in Delhi, India.
1Column percentage.Bold values are statistically significant at α = 0.05.
Logistic regression for select demographic and behavioural correlates (p < .1) of BV.
Discussion
Reproductive-age group women share the highest burden of BV-associated adverse effects on sexual health and pregnancy outcomes. The participants of the present study were recruited from two reproductive health-related OPDs frequented by women with sexual or reproductive health issues. Among them, 24.4% were diagnosed with Nugent-BV, comparable with prevalence rate in another primary healthcare centre in rural North India. 2 While the highest number of Nugent-BV positive women were between 21 and 30 years; overall age and other social factors including marital status, duration of marriage or gravidity did not significantly influence the positivity rate of this study cohort.
The association between BV and increased susceptibility to upper genital tract infections, STIs, and its implications on infertility associated with PID or endometritis have been previously documented. 17 BV has also been independently linked to tubal infertility and idiopathic infertility. Meta-analyes of studies assessing the importance of vaginal milieu in women seeking fertility treatments report 3.3 times risk for BV among infertile women with varying prevalence.18,19 Consistent with these findings, a significantly higher incidence of Nugent-BV was observed among women with infertility in this study, independent of the type of infertility (primary or secondary) or its duration.
While the role of BV among pre-menopausal women has been reiterated frequently, post-menopausal women are also susceptible to unfavourable vaginal microbiota due to decreased levels of oestrogen and progesterone and subsequent depletion of resident lactobacilli. Taylor-Robinson et al. identified as many as 35% of women attending a post-menopausal clinic to have dysbiotic vaginal flora, and 18% of women with BV. 20 The small proportion of post-menopausal participants did not provide an ideal representation of this group in this study and no significant difference in incidence of Nugent-BV was noted.
Demographic factors involved in the acquisition of BV are strongly driven by socio-culturally influenced sexual behaviours and hygiene practices. Several studies from United States and United Kingdom have identified ethnicity, increasing lifetime sex partners and higher frequency of douching as BV-associated risk factors.21–23 In the present study setting, monogamous marriages were more common, and therefore, the influence of number of partners or recent partner change on incidence of Nugent-BV could not be ideally investigated. Also, no significant association between the increasing frequencies of sexual intercourse with the risk of Nugent-BV was observed.
The role of contraception in BV has also been widely acknowledged, however with conflicting results. Among various contraceptive methods, the use of copper IUD has been strongly associated with diverse vaginal microbiota and almost three fold increased odds for BV.14,22,25 In contrast; a few other studies including ours lacked evidence to establish any effect of IUD use on Nugent-BV.12,26 This could possibly be due to the infrequent use of IUD in this study group. On the other hand, we noted in agreement with previous findings that regular condom use reduces the risk of Nugent-BV, suggesting their importance in preventing incidental perturbations of the vaginal flora.14,26–28 Moreover, consistent and correct condom use are highly effective in preventing transmission of STIs such as chlamydia, gonorrheae and HIV, thereby reducing the use of broad spectrum antibiotics and subsequent antibiotic-induced fluctuations of the resident microbial flora. Additionally, key BV associated bacteria Sneathia amnii has been identified in male non-gonococcal urethritis (NGU). 29 Also, a low prevalence of BV among oral contraceptive (OCP) users has also been observed by a few investigators; however, no effect of OCP use on Nugent-BV was observed in this study.14,26,27
Recent sexual intercourse has been linked to increased likelihood of BV, as seminal exposure not only elevates the vaginal pH but may also introduce G vaginalis, a key marker for dysbiosis. 30 Presence of G vaginalis in male partners of BV affected women was reported by Piot et al. suggesting the sexual transmission of BV. 31 Consequently, all women in this study who reported unprotected sex within 72 h preceding their hospital visit, tested positive for Nugent-BV. The direct impact of sexual intercourse on incidence of BV may, however, present a certain degree of ambiguity as transient fluctuations of the vaginal microbiota may or may not persist to establish infection. 8 Besides, this study group comprised of a high proportion of women actively trying for conception, suggesting a higher cumulative rate of sexual activity, however without significantly being associated with Nugent-BV. Therefore, in the case of history of recent sexual activity, a repeat swab may be recommended to confirm the diagnosis.
Menstrual and personal hygiene behaviours impact the occurrence of BV. Consistent with previous findings, women who used disposable menstrual protection pads were at reduced odds for Nugent-BV. 13 Contrarily, frequent vaginal douching negatively impacts the natural balance of the vaginal flora and is almost twice as likely to be associated with BV. 24 While douching is uncommon among Indian women, the use of vaginal hygiene washes was infrequent among the study participants and was not associated with incidence of Nugent-BV.
Lastly, evaluation of clinical symptoms of vaginitis revealed no significant correlation with Nugent-BV, thereby eliminating their role as potential predictors of infection. An evident limitation is the possible over-reporting of symptoms since the information was collected directly from patients in the form of questionnaires.
In conclusion, the present study has demonstrated the positive association between infertility and recent sexual intercourse with Nugent-BV, thereby underscoring the importance of regular screening, especially among women with fertility issues. The strong protective effect of condoms against Nugent-BV highlights the need to emphasise and encourage the use of barrier contraception in prevention of dysbiosis and its adverse consequences on the sexual and reproductive health of women. A welcome advantage of this would be prevention of STIs.
Footnotes
Acknowledgements
The authors acknowledge Indian Council of Medical Research, Government of India for research fellowship award (5/3/8/35/ITR-F/2019-ITR) to Ms. Apoorva Challa.
Declaration of conflicting interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt the following financial support for the research, authorship, and/or publication of this article: The work was supported by grants from Intramural Collaborative Research Award (Grant Number: 8-06/AC-06/2018-RS) by All India Institute of Medical Sciences, New Delhi, India.
Research ethics
The study was approved by the Institute Ethics Committee, All India Institute of Medical Sciences, New Delhi (IECPG-415/30.08.2018, RT-24/27.09.2018).
