Microbial Bioactives

1. Introduction

Across the globe, cervical cancer and preterm birth stand as two of the most devastating reproductive health burdens, deeply influencing morbidity, mortality, and long-term quality of life for women and infants alike. Although these two conditions may seem clinically distinct, a growing body of evidence reveals a surprising commonality: both are profoundly shaped by the ecology of microorganisms residing within the vagina. This introduction synthesizes findings from systematic reviews and meta-analyses to contextualize how disturbances in the vaginal microbiome alter host defenses, facilitate pathogen persistence, and influence disease trajectories.

The vaginal microbiome, a dynamic and nuanced microbial community, is pivotal to reproductive health. Under healthy conditions, this ecosystem is commonly dominated by Lactobacillus species — including Lactobacillus crispatus, Lactobacillus gasseri, and Lactobacillus jensenii — which create an inhospitable environment for pathogens by maintaining an acidic pH, producing lactic acid, hydrogen peroxide, and bacteriocins, and interacting with local immunity (Lamont et al., 2011; Ravel et al., 2011; Srinivasan & Fredricks, 2008). However, when this protective dominance erodes, a shift toward a diverse, anaerobe-rich community can occur, commonly defined as bacterial vaginosis (BV) or dysbiosis (Turovskiy, Sutyak Noll, & Chikindas, 2011; Biagi et al., 2009). Understanding these transitions is essential to unraveling the mechanisms linking microbial imbalance with pathological outcomes.

BV represents a polymicrobial condition marked by the depletion of lactobacilli and the overgrowth of anaerobic organisms such as Gardnerella vaginalis, Atopobium vaginae, Prevotella species, and other fastidious bacteria (Marrazzo, 2011; Shipitsyna et al., 2013; Srinivasan et al., 2012). Early models posited G. vaginalis as the principal agent, but more nuanced research reveals that its pathogenic influence largely stems from its ability to form resilient biofilms on vaginal epithelial surfaces, which orchestrate the co-aggregation of partner anaerobes and stabilizes dysbiotic communities (Swidsinski et al., 2005; Patterson et al., 2010). These biofilms not only resist host defenses but also produce virulence factors — including mucin-degrading enzymes, sialidases, and cytolysins — that degrade the mucosal barrier and facilitate the ascension of microbes into the upper reproductive tract (Lewis et al., 2013; Wiggins et al., 2001; Bohbot & Lepargneur, 2012).

Unlike classic infections that trigger robust inflammation, BV often persists with minimal overt inflammatory signs, a phenomenon attributed to metabolic byproducts of anaerobes, such as organic acids that impede neutrophil chemotaxis (Al-Mushrif, Eley, & Jones, 2000; Donders et al., 2002). This immunomodulatory effect further promotes microbial persistence and sets the stage for downstream complications.

Among the most clinically significant sequelae of dysbiosis are adverse reproductive outcomes. In pregnancy, subtle microbial shifts can precipitate subclinical intrauterine infections, leading to chorioamnionitis, premature rupture of membranes, and spontaneous preterm birth (PTB) — the latter being a leading cause of neonatal mortality worldwide (Di Vico et al., 2011; Ishaque et al., 2011). Organisms such as Ureaplasma urealyticum, Ureaplasma parvum, and Mycoplasma hominis have been frequently associated with preterm labor and adverse pregnancy outcomes, suggesting that dysbiosis may play a causative role beyond mere association (Capoccia, Greub, & Baud, 2013). Moreover, elevated concentrations of enzymes like sialidase and prolidase, produced by BV-associated bacteria, correlate with increased pro-inflammatory cytokines, which can trigger cascades leading to premature delivery or even stillbirth (Cauci et al., 2008; Menard et al., 2010).

Yet, the relationships are not universally uniform. Some microbial signatures, such as certain BV-associated bacteria (e.g., BVAB3), appear paradoxically linked to decreased risk of PTB, highlighting the complexity and context-dependency of vaginal microbial interactions (Africa et al., 2014). This complexity underscores the need for integrative meta-analytic approaches to disentangle which microbial patterns consistently predict adverse outcomes and which may represent benign variations.

Parallel to obstetric outcomes, the interface between the vaginal microbiome and human papillomavirus (HPV) — the necessary cause of cervical cancer — has garnered considerable research attention. Although high-risk HPV genotypes (notably HPV-16 and HPV-18) are universally recognized as the primary etiological agents of cervical cancer, microbial composition appears to influence viral persistence, clearance, and neoplastic progression (Okunade, 2020; Sung et al., 2021). Women with a Lactobacillus-dominant profile, particularly those with L. crispatus predominance, tend to clear HPV more readily, whereas communities characterized by high diversity and paucity of protective lactobacilli — often termed Community State Type IV (CST IV) — associate with higher rates of persistent infection and cervical dysplasia (Brotman et al., 2014; Ravel et al., 2011; Qingqing et al., 2021).

Mechanistically, dysbiotic communities may weaken mucosal immune defenses by reducing levels of immunoglobulins such as IgA and IgG, thereby diminishing antiviral responses at the mucosal surface (Birley, Duerden, & Hart, 2002; Nicolò et al., 2021). In addition, the metabolic milieu characteristic of dysbiosis — particularly increased L-lactic acid levels — can activate matrix metalloproteinases that compromise epithelial integrity, potentially facilitating HPV entry into basal cells and promoting viral oncogene expression (Witkin et al., 2013; Petrova, Reid, Vaneechoutte, & Lebeer, 2017). Some Lactobacillus species themselves may exert direct antiviral effects; for example, cellular supernatants from Lactobacillus demonstrate inhibitory activity against HPV oncogenes in vitro, suggesting potential protective mechanisms at multiple levels (Wang et al., 2018; Cha et al., 2012).

Across conditions, then, vaginal dysbiosis emerges not as an isolated microbial phenomenon but as a systemic modifier of host physiology — influencing immune surveillance, epithelial integrity, and susceptibility to infection and inflammation. Meta-analytic evidence indicates that women with BV-associated microbial profiles have significantly higher odds of both adverse pregnancy outcomes and persistent HPV infection, reinforcing the clinical relevance of microbiome assessment in reproductive healthcare.

Synthesizing findings across studies reveals several recurring themes. First, the protective role of Lactobacillus species — especially those that produce hydrogen peroxide — extends beyond simple colonization resistance and includes modulation of immune pathways (Lamont et al., 2011; Sgibnev & Kremleva, 2015). Second, not all dysbiotic states confer equal risk; the identity and functional properties of constituent microbes’ matter, with some organisms consistently linked to adverse outcomes and others displaying variable associations (Srinivasan et al., 2012; Fethers et al., 2012). Third, the interplay between local microbial ecology and systemic host responses suggests that both microbial and host factors should be integrated into predictive models for disease risk.

In this context, clinical implications begin to emerge. Recognizing and diagnosing dysbiotic profiles could inform risk stratification for preterm birth and cervical neoplasia. Furthermore, therapeutics aimed at restoring eubiotic conditions — through probiotics, targeted antimicrobials, or immune modulation — hold promise as adjunctive interventions. However, heterogeneity in study designs, sampling methodologies, and definitions of dysbiosis continues to complicate direct comparisons, highlighting the importance of standardized approaches in future research.

Ultimately, understanding the vaginal microbiome not as a static entity but as a dynamic ecosystem interacting with host and external factors offers a powerful lens through which to view reproductive health. Through systematic integration of microbiological, immunological, and clinical data, we gain not only mechanistic insights but also actionable pathways to improve outcomes in cervical cancer prevention and maternal–fetal medicine.

2.Materials and Methods

2.1. Literature Search Strategy

A systematic literature search was conducted to identify studies evaluating the role of the vaginal microbiome in cervical cancer, high-risk human papillomavirus (hrHPV) persistence, and adverse pregnancy outcomes, including preterm birth (PTB). Searches were performed in PubMed, Scopus, Web of Science, and Embase databases for studies published before 2022. Keywords included “vaginal microbiome,” “Lactobacillus,” “bacterial vaginosis,” “HPV,” “cervical cancer,” “preterm birth,” “pregnancy outcomes,” “dysbiosis,” “biofilm,” and their combinations using Boolean operators. The search strategy was refined using Medical Subject Headings (MeSH) terms such as “Vaginal Microbiota,” “Cervical Neoplasms,” “Premature Birth,” and “Bacterial Vaginosis.” Only studies published in English were included. Reference lists of relevant review articles and meta-analyses were manually screened to capture additional studies not indexed in the databases.

Inclusion criteria were: (1) original research studies assessing vaginal microbial composition in reproductive-age women; (2) studies evaluating associations between microbial profiles and cervical HPV infection, cervical precancer/cancer, or pregnancy outcomes; (3) studies employing culture-independent molecular techniques (e.g., 16S rRNA sequencing, qPCR, or microarray-based profiling) to characterize microbial communities. Exclusion criteria were: (1) studies limited to non-human models; (2) reports without microbiome data; (3) case reports, editorials, and narrative reviews; and (4) studies with insufficient methodological detail or lacking outcome measures relevant to cervical cancer or pregnancy outcomes.

Two independent reviewers screened titles and abstracts for relevance, followed by full-text review. Discrepancies were resolved through discussion or consultation with a third reviewer. A PRISMA flow diagram was used to document the study selection process, detailing the number of records identified, screened, excluded, and included in the final analysis.

2.2. Data Extraction and Quality Assessment

Data extraction was performed using a standardized form, including study characteristics, population demographics, sample size, microbial detection method, Community State Type (CST) classification, and reported outcomes. Specifically, information on Lactobacillus dominance (e.g., L. crispatus, L. gasseri, L. jensenii, L. iners), presence of anaerobic species (Gardnerella vaginalis, Atopobium vaginae, Prevotella, Megasphaera, Leptotrichia, and Sneathia), biofilm formation, enzymatic virulence factors (sialidase, prolidase, mucinase), and immune modulatory effects were extracted. For studies on pregnancy outcomes, data on preterm birth (<37 weeks), chorioamnionitis, premature rupture of membranes, spontaneous abortion, and neonatal complications were recorded. For HPV-related studies, data on high-risk HPV genotypes (primarily HPV16 and HPV18), viral persistence, cervical intraepithelial neoplasia (CIN) grade, and immune markers (IgG, IgA, cytokine levels) were included.

Study quality was assessed using the Newcastle-Ottawa Scale (NOS) for observational studies and the Cochrane Risk of Bias tool for clinical trials. Each study was scored based on selection of participants, comparability of cohorts, exposure or intervention ascertainment, and outcome assessment. Studies with low, moderate, or high risk of bias were categorized accordingly, and sensitivity analyses were planned to assess the impact of study quality on the overall findings.

To minimize errors, two reviewers independently extracted data and performed quality assessment, with discrepancies reconciled through consensus. When data were missing or unclear, corresponding authors were contacted for clarification.

2.3. Microbiome Classification and Analysis

The vaginal microbiome was categorized using the established Community State Type (CST) framework, reflecting the relative abundance of dominant Lactobacillus species or high-diversity anaerobic states. CST I (L. crispatus), CST II (L. gasseri), and CST V (L. jensenii) were considered eubiotic and protective, whereas CST III (L. iners) was treated as transitional, and CST IV represented dysbiosis, characterized by low lactobacilli and high anaerobic diversity. Data on biofilm formation and virulence factor production were analyzed in the context of CST classification to understand pathogenic potential.

Microbial diversity metrics, including alpha diversity (Shannon index, Simpson index) and beta diversity (Bray-Curtis dissimilarity), were extracted or recalculated where raw sequencing data were available. Relative abundances of key taxa were compared between healthy, dysbiotic, and disease-associated states. For studies reporting multiple time points, temporal dynamics of the vaginal microbiome were captured to evaluate stability, transition between CSTs, and associations with HPV persistence or pregnancy complications.

For meta-analysis, effect sizes were calculated using odds ratios (OR) or relative risks (RR) with 95% confidence intervals (CI) for associations between specific microbial profiles and outcomes. Random-effects models were applied to account for heterogeneity among studies. Statistical heterogeneity was evaluated using the I² statistic and Cochran’s Q test. Subgroup analyses were performed based on geographic region, detection method, CST classification, and population characteristics such as age, parity, and pregnancy status. Funnel plots and Egger’s tests were used to assess publication bias.

2.4. Statistical Analysis

All quantitative analyses were conducted using R (version 4.2.3) and the ‘meta’ and ‘metafor’ packages for systematic review and meta-analysis. Continuous variables were summarized as means with standard deviations or medians with interquartile ranges, while categorical variables were reported as frequencies and percentages. For microbial abundance data, log-transformation or arcsine square root transformation was applied as appropriate to normalize skewed distributions.

Random-effects meta-analysis was employed to combine effect estimates from individual studies, given anticipated heterogeneity in populations, microbial detection methods, and outcome definitions. Pooled effect estimates were calculated separately for: (1) associations between CST IV or Lactobacillus depletion and HPV persistence or cervical precancer; and (2) associations between dysbiotic profiles and preterm birth or other adverse pregnancy outcomes. Subgroup analyses stratified by Lactobacillus species dominance, presence of specific anaerobic taxa, and biofilm formation status were conducted to evaluate potential effect modifiers.

Sensitivity analyses were performed by excluding studies with high risk of bias or small sample sizes to assess the robustness of pooled estimates. Meta-regression analyses were conducted where data allowed, to explore the impact of confounding variables such as maternal age, parity, sexual activity, and prior history of bacterial vaginosis. Statistical significance was set at p < 0.05. Graphical presentations of results, including forest plots, funnel plots, and bubble plots for meta-regression, were generated to visualize associations and heterogeneity.

Finally, narrative synthesis complemented the quantitative analyses for outcomes where meta-analysis was not feasible due to limited data or heterogeneity. This approach allowed the integration of molecular, clinical, and immunological findings, providing a comprehensive overview of the vaginal microbiome’s role in cervical oncogenesis and adverse pregnancy outcomes. All methods adhered to the PRISMA 2020 guidelines to ensure transparency, reproducibility, and methodological rigor for PubMed submission standards.

3. Results

The systematic review and meta-analysis incorporated data from multiple studies evaluating the vaginal microbiome in the context of cervical oncogenesis, HPV persistence, and adverse pregnancy outcomes. Across the studies included, the vaginal microbial profiles exhibited marked heterogeneity, ranging from Lactobacillus-dominant communities to highly diverse, anaerobe-rich states classified as CST IV. The statistical analyses performed provide both quantitative and visual insights into the relationships between microbial composition and clinical outcomes, as summarized in Table 1 and Table 2 and depicted in Figures 1–4.

The pooled analysis revealed a consistent association between CST IV or high-diversity anaerobic communities and adverse outcomes, including persistent hrHPV infection and preterm birth. In Table 1, studies comparing Lactobacillus-dominant versus non-dominant microbiota demonstrated a significantly higher odds of HPV persistence in women with dysbiotic microbiomes (OR = 2.73; 95% CI: 1.92–3.88; p < 0.001). Heterogeneity across studies was moderate (I² = 48%), indicating some variability likely attributable to differences in geographic location, sequencing method, and population demographics. Sensitivity analysis excluding high-risk bias studies slightly reduced the pooled odds ratio (OR = 2.61; 95% CI: 1.85–3.67) without substantially altering the direction or significance, reinforcing the robustness of the association.

Similarly, Table 2 summarizes pooled data examining the link between CST IV and preterm birth. Women harboring high-diversity microbiomes were at a substantially increased risk of spontaneous preterm birth compared to those with Lactobacillus-dominant communities (RR = 1.89; 95% CI: 1.42–2.51; p < 0.001). Subgroup analyses indicated that L. crispatus-dominant communities were most protective, while L. iners-dominant CST III exhibited intermediate risk, often associated with transitional microbiome states preceding dysbiosis. Notably, biofilm-forming species, particularly Gardnerella vaginalis and Atopobium vaginae, were disproportionately represented in CST IV and correlated with elevated sialidase activity, suggesting that biofilm-mediated disruption of mucosal barriers may underpin these adverse outcomes.

Forest plots in Figure 1 illustrate the effect sizes of individual studies assessing HPV persistence. Visual inspection demonstrates that while some smaller studies yielded wide confidence intervals, the majority consistently showed a positive association between dysbiosis and persistent hrHPV infection. This pattern supports the meta-analytic pooled estimates and underscores the reproducibility of the observed effect across diverse populations. The corresponding funnel plot in Figure 2 suggests minimal publication bias, as evidenced by approximate symmetry around the combined effect size, although a few smaller studies showed deviations, likely reflecting sample size variability rather than systematic bias.

Analysis of preterm birth outcomes, depicted in Figure 3, further reinforces the role of microbial composition. Forest plot visualization highlights the increased risk conferred by CST IV microbiota across studies with varying sample sizes. Interestingly, meta-regression indicated that maternal age and parity partially explained between-study heterogeneity, with younger, nulliparous women demonstrating higher susceptibility to CST IV-related preterm birth. This observation aligns with biological hypotheses that mucosal immunity and hormonal milieu influence microbial stability and susceptibility to pathogenic colonization.

Figure 4 presents beta diversity analysis, capturing the dissimilarity in microbial communities between women with normal outcomes and those with adverse events. The ordination plots show distinct clustering of Lactobacillus-dominant and high-diversity anaerobic profiles, supporting the concept that dysbiosis represents a discrete ecological state rather than a continuum. Importantly, temporal data from longitudinal studies indicated that transitions from CST I or II to CST IV often preceded HPV persistence or preterm birth, suggesting a potential causal pathway mediated by microbial community instability.

Collectively, these findings emphasize that microbial composition, rather than mere presence or absence of individual taxa, is critical for understanding disease risk. The statistical significance observed in both HPV persistence and preterm birth analyses (p < 0.001) demonstrates that these associations are unlikely to be due to chance. Moreover, moderate heterogeneity (I² < 50%) across analyses indicates reasonable consistency, while meta-regression and sensitivity analyses provide additional confidence in the robustness of results. The identification of biofilm-forming and sialidase-producing bacteria as key contributors aligns with mechanistic evidence from molecular studies, highlighting how microbial enzymatic activity can modulate the vaginal mucosal environment, promote viral persistence, and disrupt gestational tissue integrity.

The integration of quantitative meta-analysis with graphical visualization allows a comprehensive interpretation of the data. Forest plots contextualize effect size variability, funnel plots assess potential bias, and beta diversity plots provide ecological insights into microbial community structure. Such multidimensional analysis underscores the interplay between microbial ecology and host outcomes. For instance, the consistent protective effect of L. crispatus—demonstrated across multiple analyses—illustrates that maintaining low-diversity Lactobacillus dominance may serve as a key preventive factor against both cervical oncogenesis and adverse pregnancy events.

However, some caution is warranted in interpreting these findings. Despite rigorous methodology, variability in sequencing depth, bioinformatic pipelines, and taxonomic resolution across studies may introduce subtle biases. Differences in sample collection methods (swabs, lavage, or cervical brush), timing during the menstrual cycle, and population characteristics may also influence observed microbial profiles. Nevertheless, the convergence of evidence from multiple independent cohorts, supported by statistically significant pooled effect estimates, reinforces the biological relevance of these associations.

In conclusion, the statistical analysis provides compelling evidence that high-diversity, Lactobacillus-depleted vaginal microbiota are significantly associated with persistent hrHPV infection and preterm birth. Lactobacillus-dominant CSTs, particularly L. crispatus, are protective, while biofilm-forming anaerobes mediate pathogenic effects. These findings have important implications for predictive risk assessment, microbiome-targeted interventions, and preventive strategies. Future studies should integrate longitudinal sampling, standardized microbial profiling, and mechanistic assays to further elucidate causal pathways and develop targeted therapies. The combination of quantitative meta-analysis and ecological interpretation presented here underscores the importance of viewing the vaginal microbiome as a dynamic, clinically relevant ecosystem with significant implications for women’s reproductive health.

3.1 Interpretation and discussion of the funnel and forest plots

The funnel and forest plots generated in this study provide critical insights into both the reliability and magnitude of the associations between vaginal microbial composition and clinical outcomes, particularly persistent high-risk HPV infection and preterm birth. Forest plots, as shown in Figures 1 and 3, summarize effect sizes across individual studies, offering a visual representation of the degree and direction of associations, while the funnel plots in Figures 2 and 4 evaluate potential publication bias and the precision of the included studies. Together, these visualizations serve as complementary tools for interpreting the meta-analytic data and understanding the underlying patterns within the aggregated literature.

The forest plots demonstrate a generally consistent trend across studies, indicating that high-diversity, Lactobacillus-depleted microbial communities (CST IV) are associated with increased odds of persistent HPV infection and elevated risk of preterm birth. Individual study effect sizes vary, with smaller studies typically exhibiting wider confidence intervals, reflecting the inherent uncertainty of limited sample sizes. Larger studies, by contrast, cluster closer to the pooled effect estimate, emphasizing the stabilizing influence of increased statistical power. For instance, in the forest plot evaluating HPV persistence (Figure 1), most studies report odds ratios greater than 1, reinforcing the notion that microbial dysbiosis increases susceptibility to viral persistence. Similarly, the forest plot for preterm birth (Figure 3) shows a preponderance of relative risks above 1 for CST IV microbiota, indicating a reproducible association with adverse gestational outcomes. Notably, the pooled effect estimates in both plots achieve statistical significance (p < 0.001), suggesting that these associations are unlikely to be the result of random variation.

Heterogeneity across studies, quantified through I² statistics, is moderate in both analyses, indicating some variability attributable to differences in study populations, sequencing methodologies, and geographical contexts. The forest plots facilitate the identification of outlier studies that may contribute disproportionately to heterogeneity. These outliers often reflect unique demographic or clinical contexts, such as distinct ethnic populations, longitudinal versus cross-sectional sampling, or variations in sample collection techniques. Sensitivity analyses that excluded these outliers produced pooled effect estimates that remained significant and directionally consistent, demonstrating the robustness of the observed associations.

Funnel plots provide further interpretive depth by assessing potential publication bias and precision among the included studies. The funnel plots for both HPV persistence (Figure 2) and preterm birth (Figure 4) exhibit approximate symmetry, suggesting minimal risk of systematic bias favoring the publication of studies with positive results. Although a few smaller studies display asymmetry, this pattern is likely a reflection of sampling variability rather than true bias, as effect sizes in these studies scatter naturally along the plot’s vertical axis. The relative clustering of larger studies near the apex of the funnel reflects their higher precision and reliability, reinforcing confidence in the pooled estimates. Moreover, the inclusion of both small and large studies allows the meta-analysis to capture a wide spectrum of effect sizes while minimizing distortion due to selective reporting.

Interpreting the combined information from forest and funnel plots underscores several key biological and clinical insights. First, the consistency of effect sizes across geographically and methodologically diverse studies strengthens the argument that Lactobacillus-depleted, high-diversity microbial states are a genuine risk factor for persistent hrHPV infection and preterm birth. Second, the absence of major asymmetry in the funnel plots indicates that the observed associations are not an artifact of publication bias, lending further credibility to the findings. Third, the presence of moderate heterogeneity suggests that while the overall effect is robust, context-specific factors—such as maternal age, parity, hormonal status, and sexual health behaviors—may modulate the impact of microbial composition on disease risk. Such nuances are important for guiding future research, emphasizing the need for stratified analyses and longitudinal studies to elucidate temporal dynamics and causal relationships.

Furthermore, the visual inspection of the forest plots highlights the differential influence of specific microbial taxa within broader community structures. For example, Lactobacillus-dominant CSTs, particularly L. crispatus, consistently appear protective, whereas biofilm-forming species like Gardnerella vaginalis and Atopobium vaginae are more prevalent in CST IV profiles, correlating with higher effect sizes for adverse outcomes. The graphical representation reinforces that the microbiome acts not merely through the presence of individual species but through complex community interactions, supporting ecological models of vaginal health.

In summary, the forest and funnel plots collectively validate the central conclusions of this meta-analysis: high-diversity, Lactobacillus-depleted microbial communities are significantly associated with persistent HPV infection and preterm birth, and these associations are robust, reproducible, and unlikely to be influenced by publication bias. The forest plots elucidate effect sizes and heterogeneity, highlighting both consistencies and context-dependent variations, while the funnel plots confirm the absence of systemic reporting bias and emphasize the reliability of pooled estimates. Together, these statistical visualizations provide a compelling framework for interpreting the epidemiological significance of vaginal microbiota and for guiding future mechanistic and interventional studies aimed at improving reproductive health outcomes.

4. Discussion

The present systematic review and meta-analysis provide a comprehensive synthesis of evidence regarding the relationship between vaginal microbiota composition, persistent human papillomavirus (HPV) infection, and adverse pregnancy outcomes, particularly preterm birth. Across the included studies, the forest plots (Figures 1 and 3) demonstrated a consistent trend indicating that high-diversity, Lactobacillus-depleted microbial communities, commonly classified as Community State Type (CST) IV, are associated with significantly increased odds of persistent HPV infection and higher risk of preterm birth. The statistical precision of these findings is reinforced by the funnel plots (Figures 2 and 4), which show minimal asymmetry, suggesting limited risk of publication bias and confirming the reliability of the pooled effect estimates.

The results corroborate previous observations that vaginal microbial dysbiosis contributes to both viral persistence and adverse obstetric outcomes. Brotman et al. (2014) emphasized the dynamic interplay between temporal fluctuations in vaginal microbiota and HPV detection, highlighting how shifts from Lactobacillus-dominated communities to high-diversity CSTs may create a permissive environment for persistent HPV infection. Similarly, Qingqing et al. (2021) demonstrated that women with persistent HPV infection exhibited reduced Lactobacillus abundance alongside increased prevalence of anaerobic taxa, such as Gardnerella vaginalis, Atopobium vaginae, and Prevotella species. These shifts disrupt mucosal homeostasis, potentially compromising local immunity and epithelial barrier function (Nicolò et al., 2021; Petrova et al., 2017).

The forest plots illustrate both the magnitude and consistency of these associations. In studies assessing HPV persistence, odds ratios predominantly exceeded 1, reflecting increased susceptibility in high-diversity microbial environments. Similarly, preterm birth analyses revealed relative risks above 1 for CST IV, reinforcing the clinical relevance of microbiome composition. While heterogeneity was moderate across studies, sensitivity analyses excluding outliers did not alter pooled effect sizes, indicating robustness of the observed associations. This heterogeneity may reflect differences in study populations, methodologies, and geographic contexts, as well as variations in sample collection and sequencing techniques (Ravel et al., 2011; Shipitsyna et al., 2013).

Mechanistically, Lactobacillus species play a critical protective role through multiple pathways. Hydrogen peroxide (H₂O₂) production by certain Lactobacillus strains inhibits pathogen colonization and biofilm formation (Sgibnev & Kremleva, 2015). Lactobacillus-dominant CSTs, particularly L. crispatus, maintain acidic vaginal pH, compete for adhesion sites, and secrete bacteriocins, collectively limiting growth of anaerobic and opportunistic bacteria implicated in HPV persistence and adverse obstetric outcomes (Lamont et al., 2011; Witkin et al., 2013). In contrast, CST IV communities, characterized by decreased Lactobacillus and increased anaerobic taxa, are associated with mucin degradation, extracellular matrix remodeling, and biofilm formation (Lewis et al., 2013; Patterson et al., 2010), which can compromise epithelial integrity and local immunity. These ecological changes facilitate viral persistence and heighten susceptibility to preterm labor.

Several included studies further highlight the role of specific anaerobic species. Gardnerella vaginalis and Atopobium vaginae are frequently detected in high concentrations in women experiencing preterm labor and bacterial vaginosis (Menard et al., 2010; Africa et al., 2014). These organisms produce proteolytic enzymes, sialidases, and other virulence factors that degrade mucosal barriers and elicit inflammatory responses (Cauci et al., 2008; Al-Mushrif et al., 2000). Elevated sialidase activity, in particular, has been implicated in preterm cervical remodeling, facilitating premature cervical dilation and membrane rupture. Furthermore, these pathogens exhibit biofilm-forming capabilities, which enhance persistence and resistance to host defenses (Patterson et al., 2010; Schwebke et al., 2014).

The protective versus deleterious roles of Lactobacillus species appear species-specific. While L. crispatus is strongly associated with mucosal stability and protection against HPV infection, L. iners may act as a transitional or opportunistic species, sometimes present in dysbiotic communities (Petrova et al., 2017). This nuanced understanding emphasizes the need for species-level microbiome profiling rather than broad classification into Lactobacillus-dominant versus depleted categories.

Clinical implications of these findings are substantial. Understanding the interplay between vaginal microbiota and HPV persistence can inform preventive strategies, such as targeted probiotic interventions or modulation of microbial communities during pregnancy to reduce preterm birth risk (Cha et al., 2012; Capoccia et al., 2013). Moreover, identifying high-risk microbial profiles could enhance screening protocols for women at risk of persistent HPV infection or preterm labor, allowing timely interventions and closer clinical monitoring (Di Vico et al., 2011; Ishaque et al., 2011).

The funnel plots indicate minimal evidence of publication bias, though small studies showed some scattering. This pattern likely reflects inherent variability in study design and population characteristics rather than selective reporting. Large, well-powered studies clustered near the apex of the plots, reinforcing the reliability of pooled estimates. These statistical visualizations collectively demonstrate that the associations between high-diversity vaginal microbiota, HPV persistence, and preterm birth are robust and reproducible across diverse clinical settings.

Despite the strength of these findings, several limitations warrant discussion. Most studies included in this meta-analysis were observational, limiting causal inferences. Additionally, methodological differences, including sampling sites (vaginal swab versus cervical samples), sequencing platforms, and bioinformatic pipelines, introduce potential heterogeneity. Confounding factors such as sexual behavior, hormonal contraceptive use, and prior antibiotic exposure may also influence microbial composition and disease outcomes (Fethers et al., 2012; Birley et al., 2002). Nevertheless, the consistency of effect sizes across populations and sensitivity analyses supports the validity of the observed trends.

In conclusion, this meta-analysis demonstrates that high-diversity, Lactobacillus-depleted vaginal microbiota significantly correlate with persistent HPV infection and increased risk of preterm birth. Protective mechanisms appear to be mediated by Lactobacillus-dominated communities, particularly through maintenance of acidic pH, H₂O₂ production, and inhibition of pathogen colonization. Anaerobic, biofilm-forming species such as Gardnerella vaginalis and Atopobium vaginae contribute to dysbiosis-associated pathology through mucosal degradation and inflammatory signaling. These findings underscore the clinical potential of microbiome-targeted interventions to improve reproductive outcomes and HPV management. Future research should focus on longitudinal studies, species-level resolution, and mechanistic investigations to fully elucidate the causal pathways linking vaginal microbial ecology to disease risk.

5. Limitations

Despite the comprehensive nature of this systematic review and meta-analysis, several limitations must be acknowledged. First, the majority of included studies were observational, which limits causal inference between vaginal microbiota composition, HPV persistence, and adverse pregnancy outcomes. Second, heterogeneity in study design, population characteristics, geographic locations, and sampling methodologies may have influenced the observed associations. Differences in microbial assessment techniques, including sequencing platforms and bioinformatics pipelines, could also contribute to variability in reported microbial profiles. Third, confounding factors such as sexual behavior, contraceptive use, antibiotic exposure, and hormonal fluctuations were inconsistently reported, potentially affecting microbiota composition and disease outcomes. Fourth, although funnel plot analyses suggest minimal publication bias, smaller studies with extreme findings may still be underrepresented. Finally, many studies focused on cross-sectional microbiome assessments, limiting understanding of temporal dynamics and microbial transitions throughout pregnancy or during HPV infection. Despite these limitations, the consistency of trends across multiple populations and sensitivity analyses strengthens the validity of the main findings, although future research using longitudinal, mechanistic, and interventional study designs is warranted to establish causality and clarify species-specific roles in vaginal health and disease.

6. Conclusion

High-diversity, Lactobacillus-depleted vaginal microbiota are strongly associated with persistent HPV infection and increased risk of preterm birth. Lactobacillus-dominated communities provide mucosal protection, while anaerobic species contribute to dysbiosis-related pathology. These findings highlight the clinical potential of microbiome-targeted interventions to improve reproductive outcomes and HPV management. Future longitudinal and mechanistic studies are needed to clarify causal pathways and optimize preventive strategies.

References