Concerning BV in pregnant women, it has been related to an increased risk of preclinical pregnancy loss and miscarriage in women undergoing in vitro fertilization (IVF) during the first semester
[5][6]. BV can result in the development of various infections commonly seen in everyday practice, such as chorioamnionitis and amniotic fluid contamination, postpartum and postabortal endomyometritis, endometrial bacterial colonization, preterm delivery, postpartum fever, as well as postabortal PID
[5][7]. In addition, the presence of BV may facilitate the development of an infection associated with other sexually transmitted bacteria and viruses, such as
Chlamydia trachomatis,
Trichomonas vaginalis,
Neisseria gonorrheae, human papillomavirus, and herpes simplex virus (HSV)
[8][9][10][11]. Women diagnosed with BV are at a higher risk of recurrence
[2][12][13][14]. HSV-2 seropositivity is associated with BV while effective suppressive therapy with valacyclovir did not decrease the risk
[8][15]. Importantly, BV constitutes an important risk factor for the acquisition and transmission of HIV
[16][17][18], since patients with BV presented with an almost six-fold increased quantity of HIV shed in vaginal secretions compared to those without BV
[19][20][21][22][23]. BV may cause endocervical inflammation presenting as mucopurulent cervicitis, while it has been associated with a three-fold higher risk of vaginal cuff cellulitis and abscess development after vaginal hysterectomy
[24][25][26][27]. Furthermore, BV is also related to a sixfold increased incidence of post-cesarean endomyometritis, after adjusting for various factors, such as labor duration, time after membrane rupture, and maternal age
[28]. Nevertheless, and despite the aforementioned associations, in the majority of cases, BV remains asymptomatic, mainly occurs during the menstrual cycle, and can resolve spontaneously with or without treatment
[29][30].
2. Epidemiology, Risk Factors, and Pathophysiology of BV
BV constitutes the most common cause of abnormal vaginal discharge in women of childbearing age, accounting for 40–50% of such cases
[33][34]. During pregnancy, the prevalence of BV varies from 5.8% to 19.3% while variability in prevalence exists between different races and/or ethnicities
[6]. It remains unclear whether these findings reflect genetic, socioeconomic, or behavioral discrepancies. In a combined study including 1461 pregnant asymptomatic women, the prevalence of BV rate was 12.3%
[33]. BV occurs in 33–36% of women attending sexually transmitted disease clinics, and up to 25% of those attending gynecologic clinics
[35].
Of note, Royce et al. identified that the differences in the vaginal flora of pregnant women may depend on ethnicity
[36]. They compared black and white women during the third trimester of pregnancy and observed that, in total, 22.3% of black and 8.5% of white women developed BV
[36]. In a cohort study from 2003 to 2013 in the USA including 12,340 mother–infant pairs, with 2468 being exposed to BV and 9872 being unexposed, BV-exposed mothers were more likely to be Hispanic or Black
[37].
Risk factors for BV mainly include hormonal changes, current or prior use of specific medication such as antibiotics and immunosuppressants, and the existence of foreign bodies, such as cloth or toilet tissue in the vagina
[38]. On the other hand, the appearance of BV has not been associated with the presence of comorbidities such as diabetes or immunosuppression
[39]. BV is not considered a sexually transmitted infection, however, it has been associated with smoking as well as women’s hygiene behavior, i.e., vaginal douching
[40][41] and sexual habits, such as the number of male sexual partners
[41][42], female partners
[43], new sexual partners, age of first sexual encounter, number and frequency of sexual contacts, lack of condom use
[41][44], use of an intrauterine device (IUD), especially copper-containing
[34][45], as well as infection with HIV and other sexually transmitted bacteria
[41][46]. Male genitalia may harbor BV-associated bacteria
[47][48] while circumcision is associated with reduced risk
[49]. On the other hand, therapeutic interventions targeting the male sex partner do not prevent BV recurrence in the affected female partner
[50]. Studies evaluating therapeutic interventions among women having sex with women are currently lacking. Contraception using hormonal regimens does not appear to increase the risk and may actually prevent its occurrence
[51][52]. Black race and low socioeconomic status constitute other risk factors
[53]. Interestingly, compared to women without BV, the sexual commencement of those with BV occurs at an earlier median age
[41][43]. The prevalence of the disease is related to the menstrual cycle
[54][55]. A large cross-sectional study conducted among 53,652 rural married women in China reported that menstrual cycles of more than 35 days, less than 3 days of menstruation, dysmenorrhea, and use of an intrauterine device were associated with BV
[56][57]. In a study from Kenya, it was noted that BV prevalence decreased with increasing women’s age
[58].
In a cross-sectional study of parous women in the USA, the effects of psychosocial stress, household income, and neighborhood socioeconomic parameters on the risk of BV were investigated. Having a low income was associated with an increased prevalence of BV among African American women in a statistically significant manner, but this was not the case for White American women. Furthermore, more stressful life events were significantly associated with higher BV prevalence among both African American and White American women. Moreover, neighborhood socioeconomic status was associated with the increased BV prevalence univariately by principal components analysis among White American women, but this was not to be found significant after adjusting for individual-level risk factors
[59]. Interestingly, in another cross-sectional study conducted in Tanzania, the primary education level or below was found to be a significant predictor of BV along with an age of less than 30 years, vaginal douching, HIV infection, sexually transmitted infection, early age of initiation of sexual activity, and having more than one sexual partner in their lifetime
[41]. Another study reporting data from more than 2 million Swedish women aged 15–50 years from national registry data showed that women with a lower education level had a 46% higher risk of BV, which was reduced by about 25% when adjusting for other co-variates
[60]. Moreover, in a study from 2003 to 2013 in the USA including 12,340 mother–infant pairs, with 2468 being exposed to BV and 9872 being unexposed, BV-exposed mothers were less likely to have had a college degree compared to the BV-unexposed mothers
[37].
Although various risk factors and pathogenetic mechanisms related to the development of BV have been described, its exact etiology remains unknown. The healthy vaginal flora is inhabited by a variety of
Lactobacillus spp. (90–95% of total bacteria), such as
L. chrispatus,
L. iners,
L. jensenii,
L. vaginalis, and
L. gasseri that maintain a low pH (<4.5), produce bacteriostatic and bactericidal substances and, as a result, impede the occurrence of infection
[15]. A decrease in or absence of
Lactobacillus spp. provokes an increase in the vaginal pH that leads to an overgrowth of anaerobic Gram-negative rods, leading to the development of BV
[24]. More specifically,
Lactobacillus spp. produces several chemical byproducts, such as lactic acid, bacteriocins, and hydrogen peroxide, that decrease the vaginal pH ≤ 4.5 and create an unfavorable environment for facultative pathogens
[61]. This alteration in the concentration of
Lactobacillus and pH value may facilitate an overgrowth of anaerobes, producing larger amounts of proteolytic carboxylase enzyme
[62]. This can dissect vaginal peptides into a variety of amines that are volatile, malodorous, and associated with increased vaginal transduction and squamous epithelial cell exfoliation
[62]. Interestingly, the mucinase and sialidase levels of vaginal fluid were considerably higher in women with BV compared to women with normal vaginal flora
[62]. According to a conceptual model for the pathogenesis of BV, developed by Schwebkeet et al.,
G. vaginalis is the dominant pathogen, while other pathogens mainly act synergistically as secondary intruders
[63].
Lactobacilli may have detrimental effects on the occurrence of other infections such as trichomoniasis
[64]. On the other hand, an extravaginal reservoir of vaginal bacteria may serve as an important risk factor for the recurrent episodes of BV
[1]. Importantly, the pathophysiology of BV does not include signs of inflammation, and this is the rationale for using the term ‘vaginosis’ rather than the term ‘vaginitis’ since the principal event leading to the clinical symptoms and signs seem to be a result of bacterial dysbiosis due to microbial imbalance in the microbiota of the vagina
[65].
Since the prevalence of BV is not the same in different patient populations despite having similar exposures, and due to the higher incidence in patients of African descent, genetic studies have been performed and revealed specific genetic loci implying a role for biological pathways related to cell signaling and mucosal immunity in the pathogenesis of BV
[66]. For example, a genome-wide associated study that included women from Kenya identified an association of genes encoding for molecules involved in innate immunity, such as Toll-like receptors (TLRs), interleukin-8, TIRAP, MYD88, with specific microorganisms involved in the pathogenesis of BV
[67]. Another study in HIV-infected adolescents identified an association between specific TLR single-nucleotide polymorphisms and BV, while similar studies also exist in non-HIV-infected patients
[68][69].
Additionally, there are studies suggesting an important role for metabolomics in the pathogenesis of BV. For example, a relatively recent study showed important differences in the composition and concentrations of metabolites in patients with and without BV
[70]. More specifically, when a comparison of the metabolite profiles in cervicovaginal lavage from patients with and without BV was performed, levels of 62% in 279 metabolites were significantly different in women with BV. In particular, women with BV had lower levels of amino acids and dipeptides, as well as higher levels of amino-acid catabolites, polyamines, and of the eicosanoid 12-hydroxyeicosatetraenoic acid which is a known biomarker for inflammation
[70]. In another study aiming to identify bacterial and metabolic hallmarks for BV, increased concentrations of
Prevotella,
Atopobium, and
M. hominis were more prevalent in the vaginal fluid of women with BV. The proton nuclear magnetic resonance of vaginal fluid in women with BV and without BV also identified and quantified 17 previously unreported molecules. Changes in the levels of amines, amino acids, organic acids, monosaccharides, short-chain fatty acids, and nitrogenous bases were associated with BV. More specifically, kynurenine, maltose, and NAD(+) were the most characteristic of the non-BV status, while malonate, nicotinate, and acetate were characteristic of BV
[70].
3. Bacterial Vaginosis: Diagnosis
Along with colposcopy and microscopic examination, history serves as the mainstay of diagnosis. The Amsel criteria primarily based on microscopic findings are used and require the presence of at least three of the four following parameters: (a) thin, grayish/white, homogeneous discharge; (b) pH of vaginal fluid >4.5; (c) detection of fishy odor upon addition of KOH to vaginal fluid (positive white test); and (d) the presence of significant clue cells (defined as >20% of the total vaginal epithelial cells seen on 100× magnification on saline microscopy)
[71]. Normal pH (<4.5) excludes the diagnosis of BV. The presence of cervical mucus, blood, or sperm in the vaginal secretions can increase the pH. Notably, levels of pH above 4.5 need further investigation, and a differential diagnosis between BV, trichomoniasis, and mucosal purulent cervicitis is required since all these entities can be associated with elevated levels of pH in the alkaline side.
Although the Gram stain constitutes the gold standard for the diagnosis of BV
[72], its usage is time-consuming and demands more resources and expertise compared to Amsel criteria
[73]; the detection of three Amsel criteria correlates well with Gram stain findings
[74]. However, when the Gram stain is used for the diagnosis of BV, the sensitivity and specificity of Amsel criteria are over 90% and 77%, respectively
[75]. The Gram-stained smear is examined using Nugent criteria or Hay/Ison criteria and its sensitivity varies from 62 to 100%
[76]. Nugent criteria are used to quantify or grade
Lactobacilli,
Bacteroides/
Gardnerella, and
Mobiluncus in order to create a scale of flora deviation, varying from normal (score = 0–3), to intermediate (score = 4–6) and frank (score = 7–10) BV
[61]. Of note, the Papanicolaou smear is considered unreliable for the diagnosis of BV, with a sensitivity of 49% and a specificity of 93%
[77]. Finally, the amino test is characterized by high specificity (up to 90%) and low sensitivity
[75].
Vaginal cultures of
G. vaginalis play no role in the diagnosis of BV since it is not specific to the entity and BV constitutes a polymicrobial infection; associated bacteria can be cultured even in asymptomatic women. Similarly, a cervical Pap test is of no use. Although positive cultures for
G. vaginalis are observed in the majority of symptomatic patients, this bacterium can be detected in up to 50–60% of healthy asymptomatic women, constituting its isolation not being diagnostic for BV
[62]. Multiple molecular tests as well as point-of-care tests have been made available to clinicians to facilitate diagnosis at the bedside
[78]. A relatively recent study compared the DNA hybridization test Affirm VPIII and the Gram stain using the Nugent criteria in diagnosing BV. Out of 115 positive vaginal specimens for BV as diagnosed by Gram stain, the Affirm VPIII test identified the existence of
G. vaginalis in 107 (93%)
[79]. Symptomatic women with pH changes and the presence of amine odor appear to benefit the most from this test. In another study, the combination of a positive DNA probe and vaginal pH of more than 4.5 presented a sensitivity of 95% and a specificity of 99%, respectively
[80]. The OSOM BV Blue system constitutes a chromogenic diagnostic test that depends on the activity of sialidase enzymes in vaginal fluids
[79][81]. This diagnostic modality carries a sensitivity between 88 and 94% as well as a specificity between 91 and 98% when compared with Amsel and Nugent criteria
[82]. The FemExam Test Card Uses the presence of the
G. vaginalis byproduct trimethylamine, proline aminopeptidase, and vaginal pH with a sensitivity of 91% with a lower specificity of 61%, which has been evaluated in syndromic management in resource-poor settings
[83]. With regard to molecular testing, it has been associated with superior sensitivity and specificity in detecting bacterial DNA from pathogens that have been identified are highly associated with the diagnosis of BV (i.e.,
G. vaginalis,
A. vaginae, BVAB2, or
Megasphaera type 1) and various species of
Lactobacilli (i.e.,
L. crispatus,
L. jensenii, and
L. gasseri)
[78][83][84]. It can even be performed on self-collected specimens. Various quantitative multiplex PCR assays with high sensitivity and specificity are available
[85][86][87]. Cultures and/or rapid nucleic acid testing may be helpful in differentiating BV from other infectious entities such as chlamydia or gonorrhea.
4. Post-Operative Infections after Gynecological Surgeries Associated with BV
In non-pregnant women, BV has been linked to vaginal cuff cellulitis following abdominal hysterectomy, cuff abscess, and pelvic abscess
[24][25][29]. A pelvic abscess occurs in less than 1% of women subjected to obstetric or gynecologic surgical procedures
[58]; it develops when pelvic cellulitis or pelvic hematoma expands into the parametrial soft tissue
[88]. Four studies were available concerning the risk of post-operative pelvic infections following abdominal hysterectomy among women with BV (
Table 1). The first study by Soper et al. included 161 women scheduled for abdominal hysterectomy for a benign condition
[25]. Women with BV have are at increased risk (RR: 3.2, 95% CI 1.5–6.7) for vaginal cuff cellulitis compared to those without BV
[25]. They found that women with BV had a three times higher risk of developing cuff cellulitis, cuff abscess, or both following abdominal hysterectomy compared to women without BV. Moreover, microorganisms associated with BV including
G. vaginalis,
Bacteroides spp., or
Peptostreptococcus spp. were isolated from the vaginal cuff of more than 60% of cases of cuff cellulitis.
Table 1. Post0operative infections following gynecological surgeries associated with BV.
BV: bacterial vaginosis; CI: confidence intervals; PID: pelvic inflammatory disease; RR: relative risk.
The second study by Larson et al. included 70 women undergoing abdominal hysterectomy for benign conditions excluding postmenopausal women, and it indicated a four-fold higher chance (RR: 4.1 95% CI 1.4–13.3) for the development of post-operative pelvic infection in women with BV compared to those without
[89]. Out of twenty, seven women with BV (35%), characterized by the presence of clue cells in the vaginal discharge, developed post-hysterectomy infections versus 4/50 (8%) women without BV
[89].
A nationwide study from Sweden demonstrated a 3-fold higher risk (RR 3, 95% CI 1.3–7) of post-operative infections after abdominal hysterectomy for benign conditions among women with BV compared to those without BV
[90]. Finally, Lin et al. remarked on an increased rate of post-operative infections after a major gynecologic surgery in women with BV
[91]. 175 women underwent major gynecologic surgery. These women were evaluated for the presence of BV based on Nugent’s criteria. Thirty-six percent of the positive BV women developed a post-operative fever, compared with 20% of the
Lactobacillus-predominant group of women and 12% of the intermediate-BV group (
p = 0.017). The difference between the positive-BV group and the intermediate-BV group was statistically significant (
p = 0.045 and
p = 0.007, respectively), while the difference between the intermediate-BV group and the
Lactobacillus-predominant group was not.
Women who were diagnosed with an intermediate flora were not at higher risk of post-operative infections
[91]. This could be due to the fact that they did not exclude postmenopausal women. Postmenopausal women lack
Lactobacilli in the vaginal flora and will score as intermediate flora according to the Nugent criteria. An earlier study showed that postmenopausal women had a lower rate of post-operative infection in comparison to younger women
[90].
Regarding intervention with antimicrobials, Larsson et al. noted that women with BV or intermediate flora benefit from antibiotic treatment and should be preoperatively initiated on treatment
[89]. Pre- and post-operative treatment for at least four days with metronidazole rectally considerably reduces vaginal cuff cellulitis among women with BV. Among 59 women diagnosed with abnormal vaginal flora, post-operative infections were not noted in the treated arm compared to 27% in the untreated arm. Treatment administration also decreased the vaginal cuff cellulitis rate from 9.5 to 2% among 83 women with
Lactobacillus-dominated flora. Treatment had no impact on the rate of the wound infection
[92].