Gut and Vaginal Microbiota in the Endometriosis: Systematic Review and Meta-Analysis

Endometriosis is characterized by the growth of endometrial tissue in extrauterine sites, such as the visceral and peritoneal surfaces of the pelvis and in the connective tissue of the extrpelvic region [ 1 , 2 ]. Endometriotic lesions have endometrial glands and stroma and are accompanied by cyclic and recurrent bleeding, progressive formation of fibrosis and cysts, producing chronic inflammation with increased angiogenesis, and altered immune functions [ 3 ]. Recent evidence has shown that changes in the intestinal microbiota, known as dysbiosis, may lead to the development and progression of several diseases, such as inflammatory bowel diseases, arthritis, psoriasis, neuropsychiatric diseases, and even some types of cancer [ 4 – 6 ]. These diseases are related to the inflammatory process, which is also present in endometriosis. Thus, studies investigating dysbiosis in the genital tract or pelvic cavity may be associated with the pathogenesis and/or pathophysiology of endometriosis [ 7 ]. Microbiota provides several benefits through a series of physiological functions, acting to strengthen the integrity of the intestines [ 8 ], protect against diseases [ 9 ], and also for the regulation of the immunological system [ 10 ]. However, changes in its composition can generate an imbalance in the microbiota (dysbiosis), which could predispose to the emergence of diseases and metabolic dysfunctions as the protection against microorganisms decreases [ 11 , 12 ]. Furthermore, the microbiota is also involved in adjustments to the estrogen cycle. Studies have shown that intestinal dysbiosis is a factor in increasing the levels of estrogen in the circulation, thus being able to stimulate the growth and cyclic bleeding of endometriotic lesions [ 13 ]. Therefore, this study maps the evidence in the literature regarding the association between intestinal and vaginal microbiota in endometriosis.

The study is characterized as a systematic review of the literature, written according to the PRISMA checklist (preferred reporting items for systematic reviews and meta-analyses) [ 14 ]. Registration was carried out in PROSPERO (international prospective register of ongoing systematic reviews, http://www.crd.york.ac.uk/prospero ) CRD42021260972 , 15/07/2021. POS stands for an acronym for patient, outcomes, study type [ 15 ], in which each letter represents an organized summary of the inclusion criteria presented below: P. Population of interest: women over 18 years old with and without endometriosis O. Outcomes (outcome): changes in intestinal, vaginal, and cervical microbiota S. Study type: observational studies (case-control, cross-sectional, or cohort) A comprehensive electronic search was conducted on Medline via PubMed, Embase via Elsevier, Web of Science through Web of Knowledge, Cochrane Library, and gray literature (Google Scholar) databases up to April 2022. The search strategy was organized using the following keywords: “dysbiosis,” “microbiome,” and “endometriosis” in combination with their respective synonyms, indexed terms, and free terms. Reference lists of all available primary studies were reviewed to identify additional relevant citations. The search was limited to human studies, with no language restrictions. Two reviewers (MCS and TC) independently screened the titles and abstracts through the Rayyan application ( https://www.rayyan.ai ) [ 16 ]. The potentially relevant studies were read in full, and those that met the eligibility criteria were included in this review. When there was any disagreement, a third reviewer was contacted to resolve the conflict (MIR). Studies with women over 18 years old, women with a histologically confirmed diagnosis of endometriosis, and women without endometriosis, as controls, were included. Studies with women with current inflammatory comorbidities that could alter the flora or those using antibiotics were excluded. DNA was extracted from the intestinal microbiota through stool samples collected according to the instructions of the kits used in each study. DNA was extracted from vaginal and cervical microbiota samples by swabbing, by a qualified professional. Each type of sample was processed and stored according to directions from the manufacturers of the kits used in each included study. In all samples, the 16S rRNA genomic extraction methodology was used, which is based on the detection of sequences from the highly variable region of the 16S rRNA gene, in order to identify the microbial composition and richness, that is, their taxonomic profile. In the studies included, richness was evaluated through alpha diversity, which also considers the proportion in which these species occur in each habitat (relative abundance) [ 17 ]. Two reviewers (MCS and TC), independently, extracted data using a standard form containing author, year and country of the study, age group, endometriosis diagnostic method, endometriosis staging, type of sample analyzed: intestinal, cervical, vaginal, or other, and index used to analyze the diversity and composition of the microbiota. In order to extract data from the figures, the website https://apps.automeris.io/wpd/ was used. Two reviewers (MCS and TC), independently, assessed the methodological quality (risk of bias) of the studies using the Newcastle-Ottawa scale. The results were expressed through tables and figures. Forest plots were created to illustrate the effect sizes studied for certain outcomes. For the analyses, a 95% confidence interval was used using RevMan software (version 5.4). Study heterogeneity was determined using the I 2 statistic, from 0% to 40%: may not be important; 30% to 60%: may represent moderate heterogeneity; 50% to 90%: may represent substantial heterogeneity; 75% to 100%: considerable heterogeneity [ 18 ]. When heterogeneity was suspected, random-effect model estimates were used. When no heterogeneity was observed, the Mantel-Haesnzel fixed-effect model was used.

A total of 443 studies were found across databases, and 11 studies were removed as duplicates. Thus, 432 studies were screened by reading titles and abstracts in Rayyan, and a total of 407 studies were excluded for not meeting the inclusion criteria. A total of 29 studies were selected for full text reading; 13 were excluded for the following reasons: two studies due to the wrong population including infertile women and not only due to endometriosis; ten studies due to study design: four literature reviews, two letters/comments to the editor, and four intervention studies; and one study for not presenting data from the results of the control group. Thus, a total of 16 studies were included in this systematic review. Figure 1 presents the flowchart of the study process. One cohort study [ 19 ], one cross-sectional study [ 20 ], and fourteen case-control studies were included [ 21 – 34 ], totaling 1151 women studied, of which 556 had endometriosis and 595 were part of control groups. Endometriosis diagnosis was performed through laparoscopy and stage classification according to the American Society of Reproductive Medicine (r-ASRM) [ 35 ]. The characteristics of the included studies are described in Table 1 . Ten studies [ 19 , 21 , 22 , 25 , 27 , 28 , 31 – 34 ] performed the analysis of alpha diversity using the Shannon index, which relates the richness of operational taxonomic units (OTU) and equity by the total number of species observed. A meta-analysis was performed for the intestinal and vaginal microbiota. Samples of cervical, endometrial microbiota, and peritoneal fluid were evaluated but showed heterogeneity greater than 80%; thus, the data regarding alpha diversity from these studies are available in Table 2 . (1) Intestinal Alpha Diversity. Four included studies [ 19 , 27 , 28 , 32 ] performed alpha diversity analysis of intestinal microbiota. No significant differences were found when comparing women in the endometriosis group with the control group for changes in the intestinal microbiota (SMD = −0.28; 95% CI = −0.70 to 0.14; P = 0.19; I 2 = 52%; four studies, 357 participants). The results of intestinal alpha diversity by Shannon index meta-analysis are represented in Figure 2 . (2) Vaginal Alpha Diversity. Two studies analyzed the vaginal microbiota [ 19 , 22 ]. When comparing the richness of diversity, no significant difference was found between microbiota samples from women with endometriosis and women from the control groups (SMD = −0.68; 95% CI = −1.72 to 0.35; P = 0.19; I 2 = 66%; two studies, 49 participants). The results of vaginal alpha diversity by Shannon index meta-analysis are represented in Figure 3 . Five studies [ 22 , 25 , 27 , 32 , 34 ] performed the analysis of alpha diversity using the Simpson index, evaluated to measure community diversity in order to represent the proportion of species in a sample, for example, the dominance of one species among others, and to measure differences in diversity between populations. A meta-analysis was performed for intestinal and peritoneal fluid. (1) Intestinal Alpha Diversity. Two studies [ 27 , 32 ] analyzed the intestinal microbiota. When comparing endometriosis and the control group, no significant difference was found between microbiota samples (SMD = −0.14; 95% CI = −0.63 to 0.34; P = 0.56; I 2 = 0%; two studies, 65 participants). The results of intestinal alpha diversity by Simpson index meta-analysis are represented in Figure 4 . (2) Peritoneal Fluid Diversity. Peritoneal fluid analysis was evaluated in three studies [ 25 , 32 , 34 ], but data from Yuan et al., [ 34 ] were in the median and confidence interval. Thus, two studies [ 25 , 32 ] were included in the analysis. Although the groups with endometriosis showed higher values, no significant difference was found between microbiota samples from peritoneal fluid (SMD = 0.34; 95% CI = −0143 to 0.81; P = 0.17; I 2 = 42%; two studies, 131 participants). The results of peritoneal fluid diversity by Simpson index meta-analysis are represented in Figure 5 . The considerable heterogeneity found between the studies can be attributed to different factors that change the characteristics of the microbiota, such as diet, use of medication (especially antibiotics), country of residence of the studied population, cultural factors, as well as lifestyle habits. Due to the wide variety of presentations of the results of the microbiota compositions, we were not able to perform a quantitative analysis. Thus, the main results found in studies regarding the composition of the microbiota are presented in a descriptive way. Information regarding the methodological aspects and results is described in Table 3 . The methodological quality assessment of the included studies was performed using the Newcastle-Ottawa scale (NOS). NOS were used to assess a cohort study (one study), case-control studies (14 studies), and a cross-sectional study (one study). For case-control studies, in all studies, a quality mean of 8/9 was obtained. We consider studies with 7 stars or higher as high quality and have a low risk of bias. For the 6/9 cohort study, high risk of bias, and for the cross-sectional study, the quality score was 7/9, a high quality and low risk of bias. Table 4 shows the results of the methodological quality analysis.

In this systematic review, four meta-analyses were carried out regarding alpha diversity assessed by Shannon and Simpson indexes in the intestinal, vaginal, and peritoneal fluid samples, with no significant association. However, when analyzing each study individually, differences were noted between the microbiota of women with endometriosis compared to the microbiota of women in control groups. This occurred in the analysis of intestinal and vaginal samples. Simpson's index represents the proportion of species in a sample, measuring the diversity of the community. The range is from 0 to 1, where high scores indicate high diversity. Analyzing the studies included in the analysis of the intestinal sample, the control population has greater diversity by the Simpson index, while in the peritoneal fluid, this diversity is greater in the population with endometriosis, although showed no significant difference, possibly due to the limited number of studies. The intestinal dysbiosis process may be related to the pathogenesis of endometriosis, as it compromises the barrier function, causing increased permeability and entry of microbial metabolites, which may trigger inflammatory changes, leading to acute inflammation in places outside the gastrointestinal tract, such as the peritoneum [ 36 ]. This change could result in low-grade inflammation and, over time, lead to changes in the performance of macrophages and their ability to phagocytose newly implanted endometriotic lesions [ 37 ]. It is then suggested that the microbiota, particularly in a state of dysbiosis, may contribute to immune activation that favors peritoneal inflammation and possibly the progression of endometriosis. The study by Wang et al. [ 29 ], included in this review, evaluated the presence of inflammatory markers in the peritoneal fluid. After the analysis, it was detected that the levels of IL-6, IL-10, IL-13, and TNF- α were significantly higher in women with endometriosis and infertility ( P < 0.05). In order to survive, endometriotic implants need a blood supply, and this angiogenesis process is regulated by factors such as vascular endothelial growth factor. In peritoneal fluid, vascular endothelial growth factor is produced mainly by macrophages, and its expression is directly regulated by estradiol and progesterone. TNF- α and IL-8, also secreted by peritoneal macrophages, are other potent inducers of angiogenesis and lesion proliferation. The role of TNF- α in stimulating endometrial cell adhesion and inducing angiogenesis is particularly needed in the early stages of the onset of endometriosis [ 38 , 39 ]. One of the theories about the origin of ectopic endometrial tissues is the “retrograde menstruation theory”, in which the reflux of menstrual debris with viable endometrial cells into the pelvic cavity occurs, resulting in adhesion to peritoneal surfaces and proliferation [ 40 ]. Garcia-Valesco and Arici [ 41 ] have shown that in an inflammatory peritoneal environment, this adhesion is more intensified. Khan et al. [ 42 ] propose the hypothesis of bacterial contamination. The authors suggest that a substantial amount of endotoxin in the peritoneal fluid, related to menstrual blood reflux, would be involved in pelvic inflammation and promote Toll-like receptor 4- (TLR4-) mediated growth and progression of endometriosis. Thus, lipopolysaccharide (LPS), a component of the outer membrane of gram-negative bacteria, can be the initial trigger and bacterial contamination its source in the intrauterine environment, and it could be the primary cause in the regulation of endometriosis growth, alone or in combination with ovarian steroids [ 42 ]. Most studies included in this systematic review showed an increase in gram-negative bacteria in women with endometriosis in either intestinal, cervical, or vaginal samples [ 19 , 21 – 23 , 25 , 26 , 29 , 33 ], corroborating Khan et al.'s [ 42 ] hypothesis. Another hypothesis in which the intestinal microbiota could be related to endometriosis is through the regulation of estrogen. This process would be linked to estrobolome, defined by Plottel and Blaser [ 43 ] as “the aggregate of enteric bacterial genes whose products are capable of metabolizing estrogens.” Thus, intestinal microbial richness would regulate systemic estrogen levels through the action of β -glucuronidase, present in some species of bacteria. In the intestine, this enzyme converts hepatically conjugated estrogens, which would otherwise be excreted to their active form, leading to their reabsorption [ 44 ]. Therefore, in a state of dysbiosis, there would be an increase in systemic levels of estrogen, which can be transported to distal mucosal sites, such as the endometrium [ 44 ]. One of the bacteria that presents ß -glucuronidase activity is the Escherichia coli species, a bacterium belonging to the phylum Proteobacteria . In our review, the studies by Nabiel et al. [ 26 ], Khan et al. [ 23 ], and Akiyama et al. [ 21 ] showed a significant increase in E. coli species in women with endometriosis. Besides, the study of Chang et al. [ 33 ] indicated the presence of Proteobacteria members as one unique population in the cervix of women with endometriosis. A dysregulation of the balance between the ratio of bacteria belonging to the Firmicutes and Bacteroidetes phyla is presented in women with endometriosis and may be acting in the deregulation of estrogen metabolism since the bacteria in these phyla have genes related to glucuronidase [ 45 , 46 ]. The increase in bacteria belonging to the phylum Firmicutes was presented in studies by Khan et al. [ 23 ], Hernandes et al. [ 22 ], Lee et al. [ 25 ], Shan et al. [ 27 ], and Chang et al. [ 33 ], that last one also indicating the increase in Firmicutes community in infertile patients as compared to fertile patients, especially Lactobacillus. Analyses of studies included in this systematic review indicate that another aspect to be considered is the relationship between microbiota and the stage of endometriosis. As evidenced by Ata et al. [ 19 ], highlighting the complete absence of Atopobium (belonging to the phylum Actinobacteria ) in the vaginal and cervical microbiota of the endometriosis group in stages 3 and 4. Furthermore, in the cervical microbiota, Gardnerella , Streptococcus , Escherichia , Shigella , and Ureoplasma were increased in stages 3 and 4, all containing potentially pathogenic species. The results of the study by Perrotta et al. [ 20 ] also suggested that the vaginal microbiota can predict disease stage since an OTU within the Anerococcus genus differed significantly in abundance between women in stages 1 and 2 and stages 3 and 4 of endometriosis during the menstrual phase. Cregger et al. [ 47 ] identified 56 significantly different OTUs on the day of surgery between the stage 3 patient with endometriosis and all others analyzed. Suggesting that changes in the bacterial community level may be indicative of severe active endometriosis and provide an additional explanation for the decrease in pregnancy rates in these women, suggesting that the analysis of the bacterial community profile could help in the diagnosis of endometriosis in asymptomatic infertile women. Moreno et al. [ 48 ] suggested that non- Lactobacillus -dominant compositions may trigger an inflammatory response in the endometrium, which may be an explanation for their association with negative pregnancy outcomes. Nine studies included in this systematic review [ 19 , 21 – 24 , 26 , 31 – 34 ] evaluated Lactobacillus spp. in cervical, endometrial, or vaginal samples; however, due to differences in the presentation of results, it was not possible to determine whether a difference occurred between women with endometriosis and women in control groups. Leonardi et al. [ 7 ] also performed a systematic review of the endometriosis-microbiome interaction in human and animal studies. The results found in our systematic review and meta-analysis corroborate the results found in the systematic review by Leonardi et al. [ 7 ], in which the authors stated that endometriosis seems to be associated with an increased presence of Proteobacteria , Enterobacteriaceae , Streptococcus spp., and Escherichia coli , and also they indicated an association between the increase in the phylum Firmicutes and the genus Gardnerella . However, our review differs from theirs as our assessment has included studies only with humans, including intestinal, vaginal, cervical, endometrial, and peritoneal fluid microbiota. Recently, a literature review published by Talwar et al. [ 49 ] carried out a synthesis of the processes involved in the pathogenesis of endometriosis and its relationship with the intestinal microbiota, indicating that a reduced gut microbiome diversity and an elevated Firmicutes/Bacteroidetes ratio have mostly been linked with increased endometriosis risk, contributing to highlight our results. Similarly to our study, Talwar et al. [ 49 ] demonstrated the results of studies evaluated in our systematic review [ 19 , 20 , 27 , 28 , 32 ], presenting their strengths and limitations. In addition to that, our systematic review contributes by bringing the quality assessment of these studies and the quantitative analysis of the results through meta-analysis. Most of the studies included in this review analyzed the intestinal microbiota, which was to be expected, since studies on microbiota started by evaluating the gastrointestinal tract. In recent years, other organs and fluids have also begun to have their microbiota evaluated. In our study, the cervical, vaginal, endometrial, and peritoneal fluid microbiota were also analyzed. These new analyses become important for clinical practice, enabling a greater understanding of the disease and the search for new forms of treatment for women with endometriosis. The analysis technique used in the included studies was the 16S rRNA genomic extraction methodology. However, in recent years, there has been an improvement in culture methods, defined as “culturomics”, consisting of the use of different culture conditions, such as the use of selective liquid and/or solid media for microorganism growth, variable temperature, and time of incubation [ 50 , 51 ]. Metagenomics and culturomics are different analytical methods that pursue the same objective: the detection of as many microbial species in a particular ecosystem as possible. As a strong point, the possibility of analyzing a high number of samples influences the speed of analysis results [ 51 ]. On the other hand, it is still a technique that presents limitations, such as the efficacy of DNA extraction is a crucial operator-dependent step that could affect the repeatability of the test [ 50 ]. This shows that although we are making progress in microbiota research, we still have limitations to overcome. The results presented here indicate that in women with endometriosis, there are changes in relation to the species/genera of bacteria in the microbiota. These findings show the importance of including the analysis and management of the microbiota in clinical practice, as well as the importance of studies in this area. As limitations to this study, it is highlighted that the microbiota is directly influenced by lifestyle, socioeconomic, cultural, environmental, and genetic factors. Therefore, primary studies should search for ways to minimize these interferences or find ways to incorporate them as a measurable factor. In addition to these factors, there are also limitations in the methodologies of the studies, with no standardization regarding which phase of the menstrual cycle the samples are collected, hormonal levels, and the presentation of the results of the analyzed species, resulting in substantial heterogeneity.

Alpha diversity was not significantly different for intestinal and vaginal microbiota in women with endometriosis compared to women without endometriosis. However, when analyzing the genus/species in the composition of the microbiota in women with endometriosis, we found a possible relationship that must be further investigated.