Study inclusion

A literature search on PubMed, Google Scholar, Semantic scholar, Crossref, and Web of Science databases yielded 1,979 articles. After deduplication, 1,280 studies were removed. Further screening based on title and abstract reading eliminated 640 articles. Full-text screening of the remaining 59 articles excluded 22 studies (14 reviews and 8 papers not relevant to the review’s topic). Ultimately, 37 studies were definitively selected for this review (Fig. 1; Table 1).

Table 1 Study characteristics and outcomes investigatedClassification of the included studies

The 37 studies included in this review originated from various countries: 22 (59.5%) were from the USA, three (8.1%) were from Australia, Canada, and the Netherlands, two (5.4%) were from Brazil, and one each from Argentina, Japan, Kenya, and South Africa. The studies were either retrospective (n = 20) or prospective (n = 16), with one study not providing this information.

The studies encompassed 7,380 childbearing women, of which 3,641 (47.9%) were of African and African American ethnicity, 2,329 (31.6%) were of White-Caucasian origin (including European, North-African, and Turkish origins), 516 (7%) were of Asian and Pacific Islander origins, 387 (5.2%) were of Hispanic and American Indian origins, 178 (2.4%) were of multiracial origin, and no racial information was provided for the remaining women (n = 339 (4.6%)).

Influence of personal hygiene on vaginal microbiota

Personal hygiene could influence the stability of the vaginal microbiota. In the literature, vaginal douching or other intravaginal practices (IVP) (such as intravaginal cleansing or insertion of products) represented the main studied practices. A wide variety of these intimate hygiene practices are influenced by religion, social and cultural traditions, or patients’ education. Thus, three papers reported the effect of personal hygiene on vaginal microbiota [16,17,18].

The study by Van Der Veer et al. evaluated the effect of a commercial douche on the vaginal microbiota in 25 healthy Dutch women (median age 24 years (22–29)), of whom 60% used combined oral contraceptives. The authors found that the composition of the vaginal microbiota and vaginal pH were not affected by vaginal douching, while menstruation was reported as an influencing factor involved in modifying the vaginal microbiota, doubling the presence of anaerobes frequently associated with vaginosis (OR = 1.7 with a 95%CI [1.0-2.8]). Similarly, douching during menstruation significantly increased the risk of dysbiosis with an OR of 2.6 times (OR = 2.6 with a 95% CI [1.0-6.5]), compared to patients without menstruation (p = 0.099). Moreover, vaginal douching appeared to promote Candida albicans infections [18], probably due to a proinflammatory vaginal immune environment.

Another study evaluated the impact of IVP among Kenyan women aged 18 to 45 years at risk of HIV acquisition. No significant difference was observed in the composition of their vaginal microbiota, even three months after IVP cessation. However, both studies are difficult to generalize to the entire population as they were conducted at single center (Netherlands and Kenya) with limited sample sizes (25 and 58 patients, respectively). Additionally, the 3-month follow-up duration might be too short, and the cessation of IVP might have been overestimated [17, 18].

A recent observational study conducted on 33 sexually active women revealed that douching cessation alone was not associated with a modification of the vaginal microbiota or an increased risk of bacterial vaginosis onset. Other extrinsic factors (such as antibiotic use, lubricants, diet, smoking cessation or condom use) would influence the vaginal microbiota much more [19].

Influence of hormonal contraception on vaginal microbiota

A 2019 United Nations report showed that approximately 407 million women worldwide used hormonal contraception, with 17% using intrauterine devices, 16% using oral contraceptive pills (OCP), 8% using injectable contraceptives, and 2% using an implant [20]. Since the vaginal microbiota is altered by the balance between estrogen and progesterone [21], it was likely that hormonal contraception could also influence changes between CSTs present in the vagina and colonization by Lactobacillus [22].

Studies evaluated various contraceptive solutions used systemically or non-systemically, including OCP, an injectable contraceptive and a levonorgestrel intrauterine system (LNG IUS). They revealed that OCP played a significant protective role in the vaginal microbiota [21, 23,24,25]. Indeed, among women using this method of contraception, the presence of H2O2 producing-lactobacilli was much more significant, with a vaginal microbiota mainly belonging to CST type I and much lower risks of bacterial vaginosis. Women using depot-medroxyprogesterone acetate (DMPA), an injectable contraceptive, also exhibited a predominance of CSTs containing predominantly lactobacilli, although these bacteria species produced much less H2O2 compared to what had been observed for OCP [26,27,28]. In these patients using DMPA, a high presence of Atopobium vaginae and/or Prevotella bivia was observed, indicating the presence of bacterial vaginosis at the time of sampling. Evaluation of the LNG IUS showed similar results with an increase in bacterial vaginosis and the predominance of CSTs less populated by Lactobacillus, thus tending towards a non-optimal vaginal microbiota. These results were obtained from two studies, a large cohort of 682 women, allowing for the control of confounding factors (including birth control, ethnicity, education, employment, health habits, dietary habits, sexual history, and the women age), and a second one with a small sample size (n = 11), suggesting an obvious lack of statistical power [24, 25]. However, since these two studies observed similar results, it seems likely that LNG IUS could significantly influence the composition of the vaginal microbiota. Furthermore, in the large cohort conducted by Brooks et al., the use of combined OCP with the intrauterine system favored the presence of H2O2-producing Lactobacillus in these patients compared to women exclusively using the intrauterine system [24].

However, studies comparing LNG UIS to copper intrauterine devices (CIUD) showed that the former device was more commonly associated with an increase in microbiota belonging to CST type I and greater stability of the vaginal microbiota after insertion of the LNG IUS placement compared to CIUD. As a result, no statistical difference was found. This is likely due to the lack of statistical power in the included studies, which can be attributed to the very limited size of the study population [24, 25, 29, 30].

In studies examining the influence of lifestyle on the vaginal microbiota, a major bias factor must be systematically considered: the ethnic origin of the patients. Indeed, this factor strongly influences the composition of the vaginal microbiota, especially depending on the hormonal contraception used. As 28.5% of included articles studied African or African American women, almost 50% of participants were of black ethnicity (Table 1). Women of African and African-American origin exhibited a much less stable vaginal microbiota over time, were more prone to STIs, and had higher risks of contracting HIV-1 [31]. Due to the high levels of hormonal contraception in this population, many more studies focused on African/Sub-Saharan women to assess the effectiveness of subcutaneous injectable contraceptives (such as DMPA) in preventing bacterial vaginosis [26,27,28]. Different studies have shown that a higher proportion of women presented a more diversified vaginal microbiota with the use of DMPA (in Sub-Saharan women according to 12 research sites) [28], without an increase in the occurrence of STIs, as previously suggested [24]. This result could be explained by a hypoestrogenic status induced by DMPA. As noted earlier, estrogens influence glycogen deposition on the vaginal wall, allowing colonization of the vaginal epithelium by Lactobacillus producing lactic acid, and thus contributing to maintaining the acidic pH of the vagina (Fig. 2). Conversely, this hypoestrogenic status could be associated with a lack of lactobacilli abundance and greater diversity of the vaginal microbiota, promoting infections [26]. Comparison of DMPA use by ethnic origin (African and African-American vs. Caucasian) among women with identical CSTs at baseline showed that Caucasians maintained CSTs dominated by Lactobacillus over time, while women of African and African-American origin had modifications in their microbiota that became more diverse, with a depletion of lactobacilli [32]. These results highlight the association between hormonal balance, physiological factors, and the composition and diversity of the vaginal microbiota. This association appears to be influenced by ethnic origin and the use of the contraceptives, particularly DMPA (Fig. 2).

Fig. 2

Effects of different lifestyle factors on the vaginal microbiota composition and their consequences. BDPE, benzo[alpha]pyrene diol epoxide; pH, vaginal pH; STI, sexually transmitted infection

Influence of sexual behavior on vaginal microbiota

A potential link between sexual activity and the vaginal microbiota was assessed in 14 articles, establishing that sexual activity can influence the composition and diversity of this microbiota and also plays a major role in lower genital tract inflammation and the occurrence of STIs [19, 33, 34]. Furthermore, certain sexual habits such as penile-vaginal sex, receptive oral sex, the use of sex toys, and condom use could also directly modify the composition of the vaginal microbiota [35].

The relationship between sexual activity and the vaginal microbiota was studied in 97 pubertal virgin patients to assess the evolution of the vaginal microbiome before and after the onset of sexual activity [36]. In this study, women who had not had sexual intercourse by the end of the study follow-up maintained a stable composition of their vaginal microbiota over time, mainly colonized by Lactobacillus species. In contrast, the vaginal microbiota of sexually active women belonged to CST type IVB with a predominance of Gardnerella vaginalis. G. vaginalis is typically found in low abundance in the normal vaginal microbiota and contributes to the protection against genital infections by maintaining the acidity of the vaginal pH. However, a high relative abundance of G. vaginalis can have adverse effects on reproductive and sexual health outcomes by increasing the risk of bacterial vaginosis [36]. Similar results were observed in studies examining the effects of sexual intercourse with a stable partner [31, 37, 38]. Indeed, a comparison of vaginal microbiota composition between women with stable partners and those with multiple partners showed no significant difference between the two populations, whereas women engaging in penile-vaginal intercourse were more likely to be colonized by a CST type III, with G. vaginalis being present as a minority [30, 38, 39]. These results suggest that this CST type III with a predominance of L. iners may not be protective and may represent a risk factor for developing STIs. Furthermore, studies suggest that G. vaginalis is dynamically acquired from the penile skin microbiota of a sexual partner. This reinforces the hypothesis that G. vaginalis can be sexually transmitted during unprotected intercourse [29, 40]. Finally, van Houdt et al. observed that having sexual activity but living separately from the partner was also associated with a modification of the vaginal microbiota and with an increased risk of infection, particularly with C. trachomatis [42].

Studies on the vaginal microbiota of women with or without bacterial vaginosis within monogamous heterosexual couples show that penile vaginal intercourse influences the composition and diversity of the vaginal microbiota, promoting inflammation of the vaginal epithelium [43]. These modifications of the vaginal microbiota occur, again as with G. vaginalis, upon contact with the penile skin microbiota of the partner during unprotected intercourse. Similarly, in sex workers, modifications of the vaginal microbiota have been observed, particularly during oro-vaginal intercourse where the presence of the Shuttleworthia genus in the vaginal microbiota is associated with an increased risk of bacterial vaginosis in this population [29, 40, 42]. Shuttleworthia is a commensal bacterial genus of the oral cavity. Its presence in the vaginal cavity is associated with a significant increase in vaginal microbiota diversity correlated with an absence (or a very significant decrease) of Lactobacillus, contributing to favoring the risk of acquiring STIs, especially HIV, among sex workers [40]. It could also be related to human papillomavirus (HPV) infections with low-grade squamous intraepithelial lesions, precancerous lesions of the cervix [44]. Further research is needed to clarify the role and impact of sexual behaviors on these vaginal bacterial communities [40, 45].

Other sexual habits, including homosexual or bisexual activity, receptive oral intercourse, or sharing unwashed sex toys, significantly increase the risk of modifying vaginal bacterial communities towards CST type IV, thus favoring infections [39, 45]. This risk could be reduced by treating both sexual partners [39, 44, 46]. Furthermore, prostitution is a significant risk factor in modifying the vaginal microbiota. It is often associated with diverse sexual intercourse, multiple partners, random condom use, and high contraceptive use [40, 47, 48]. The vaginal microbiota of sex workers is mostly lacking in Lactobacillus, with vaginal bacterial communities of type CST IV, leading to a much higher risk of contracting STIs, especially HIV [40, 48]. It is worth noting that studies on sex workers have been conducted mainly in developing countries, creating a likely analysis bias but also highlighting a major public health issue in these countries that would require the implementation of prevention plans to protect sex workers and limit the spread of STIs, particularly HIV.

Influence of psychosocial stress on vaginal microbiota

The function of the human immune system is influenced by high levels of stress, potentially increasing susceptibility to infections [18]. Although the mechanisms underlying this susceptibility are not clearly established, the vaginal microbiota appears to be altered by perceived stress (Fig. 2). This modification of microbiota could be due to the presence of biogenic amines and pro-inflammatory responses altering vaginal physiology [23]. Biogenic amines are produced by the decarboxylation of amino-acid, which involves the consumption of hydrogen ions. Decreased hydrogen ions would increase vaginal pH, completely destabilizing the balance of the vaginal microbiota, especially that of lactobacilli. This phenomenon would thus favor the vaginal colonization by a more diversified bacteria community in women exposed to psychosocial stress [17, 49]. Furthermore, stress increases adrenal corticotropic hormone released by the pituitary gland. This hormone weakens immune defenses by inducing pro-inflammatory responses, and causing cortisol secretion, which binds to beta-adrenergic receptors on immune cells, reducing their effectiveness. Cytokines, polypeptide mediators of immunity and inflammation, also play an essential role by acting as true neurotransmitters in the central nervous system during stress, affecting the immune response. Moreover, the destruction of epithelial cells during the inflammatory process caused by pH modification would prevent the deposition of vaginal glycogen, limiting the carbon, which is an essential nutrient source used by Lactobacillus species in the vaginal microbiota (Fig. 2) [18].

Among American-Indian women living with a high perceived stress level and presenting high rates of STIs, an association between the presence of CST type IV and life traumas and self-esteem issues has been observed, as well as the implication of stressful situations in the association between these CST type IV and bacterial vaginosis [51].

Women with a 5-unit increase in stress, assessed by Cohen’s 10-point Perceived Stress Scale (PSS-10), had a higher risk of developing bacterial vaginosis, associated with a low abundance of Lactobacillus and a predominance of anaerobic bacteria [51]. In these women, the risk of transitioning from CST type III to CST type IV increased by 26%. In contrast, only 19% of women with CST IV at baseline had a change in their vaginal CST evolving toward types I or II after a 5-unit increase in the PSS-10 score. This study thus confirmed the association between psychosocial stress and the modification of vaginal microbiota, often resulting in a decrease in Lactobacillus (Fig. 2) [16, 26, 51]. Even more notably, psychosocial stress has been demonstrated as a factor influencing the composition of the vaginal microbiota in > 80% of women with African or African-American origin [46, 50, 51]. However, this latter result should be moderated since we have previously seen that in this population, the percentage of women with vaginal bacterial communities of CST III and IV was significantly higher than in other populations regardless of psychosocial stress [