Association Between Polycystic Ovary Syndrome and the Vaginal Microbiome

A Case-Control Study

Xiang Hong; Pengfei Qin; Kaiping Huang; Xiaoling Ding; Jun Ma; Yan Xuan; Xiaoyue Zhu; Danhong Peng; Bei Wang


Clin Endocrinol. 2020;93(1):52-60. 

In This Article


In this study, distinct difference in vaginal bacterial structures was observed between PCOS and healthy control women, including α and β diversity aspects. The relative abundance of L. crispatus in the PCOS group was significantly lower than controls, and the relative abundance of Mycoplasma and Prevotella was significantly higher than controls. Mycoplasma could be a potential biomarker for PCOS screening, but further studies are required to confirm this observation. Sensitivity analyses showed that these results were not changed when controlling the BMI and vaginal cleanliness. It suggested that even among the samples in the same vaginal cleanliness grading, based on traditional smear examination, the vaginal microbiome was still a distinguishing factor between groups. This is the first study to identify associations between the vaginal microbiome and PCOS, describing for the first time characteristics of the vaginal microbiome in women with this disease.

Bacterially, the vaginal microbiome is a low-diversity, Lactobacilli -dominated community, where bacteria maintain vaginal homoeostasis.[9] This is particularly true for the Lactobacilli genus, which benefits the host by producing lactic acid as a fermentation product that maintains vaginal pH at 3.5–4.5.[29] However, based on current evidence, the vaginal microbiome would become altered concomitant with the female hormone changes. For example, in perinatal period or the menopause, the vaginal microbiome exhibits specific features.[29] MacIntyre et al[30] observed that vaginal microbiome composition dramatically changed postpartum, to become less Lactobacillus spp. dominant, with increased alpha diversity. Muhleisen et al[31] concluded that Lactobacilli decreased in line with menopause-related hormonal and vaginal epithelial changes. Although our study found that the vaginal microbiome changes between groups were basically stable despite of the hyperandrogenism status, we still speculate that changes in the vaginal microbiome, especially at the Lactobacillus genus level, are the microbiological outcomes of hormone disorder.

With developments in microbiome detection technology, some omics-based indices, such as α and β diversity, have been associated with disease status.[17] Although they are not widely accepted by clinicians at present, we believe that they are of scientific values and have potential for future clinical application. Meanwhile, researchers have determined that the functions of different Lactobacillus species are not the same, especially L. iners and L. crispatus, which were reported as negatively correlated in abundance level.[32] Our study observed that decreases in Lactobacillus in PCOS women were very mild, but the relative abundance of L crispatus was significantly decreased in PCOS women, suggesting L. crispatus is potentially a beneficial bacterium. Importantly, these data are consistent with other reports.[33] Equally, our study suggests that traditional vaginal secretion smear examinations may not be suitable for detecting imperceptible changes in specific Lactobacillus species. Inversely, our study found that the proportion of CST type I in PCOS women was statistically lower than healthy women. Thus, the CST classification of vaginal microbiome based on omics data seemed to be more scientific.

Whether the vaginal microbiome impacts on the PCOS occurrence or development is still unknown. Normally, disruption in 'normal' vaginal equilibrium is defined as bacterial vaginosis (BV), which is characterized by increasing pH, epithelial cell destruction and local inflammation.[34] A recent study observed that the microbiome along the female reproductive system was a continuum[15] and that vaginal microbiome disorders led to preterm or premature rupture of membranes.[12,35] Similarly, long-term chronic local inflammation is related to cervical cancer development.[10] While direct evidence is limited for bacterial function at the ovary, the potential impact of the vaginal microbiome on the female internal genital tract is worth investigating.

In recent decades, studies have investigated the Mycoplasma genus, which is the smallest free-living microorganism.[36] The most common Mycoplasma species in the vagina is M. hominis, which is regarded as an opportunistic pathogen, with a 20–50% prevalence.[36]Mycoplasma lipoproteins are potent initiators of inflammatory reactions in mycoplasma infections, activating host cells by inducing pro-inflammatory cytokine production, for example tumour necrosis factor (TNF-α), interleukin IL-1 and IL-6.[37] It has been reported that vaginal Mycoplasma infections are associated with preterm birth, spontaneous abortions and infertility issues and that retrograde infection in the genital tract may be causative reasons for these issues.[38] In a similar way, we suspect that Mycoplasma may also be associated with oviduct and ovary diseases. Except for vagina colonization, Mycoplasma has been reported in the blood, especially human immunodeficiency virus-related Mycoplasma (M. perforation, M. fermentans and M. pirum).[39,40] Associated Mycoplasma pathogenicity and pathogenic mechanisms are still unclear, and long-term Mycoplasma carriers are even associated with hepatocellular cancer.[41] Thus, the pathogenicity is systemic, including the potential ovary damage. To confound this notion, it has been shown in mouse models that female sex hormones have profound effects on the vaginal colonization of M. hominis; 90% of mice treated with oestradiol were M. hominis positive,[42] making causal correlations between Mycoplasma and PCOS more puzzling, but highlighting the importance of knowing how the hormone profiles are associated with vaginal microbiome.

Our study could not infer causal associations between PCOS and the vaginal microbiome, because of its cross-sectional design. The vaginal ecosystem is ecologically dynamic; therefore, an optimized study design would include a prospective approach with frequent sampling, including at different cycle phases, so as to fully reveal the characteristics of vaginal flora.[29] Meanwhile, it is necessary to collect vaginal microbiome before, during or after syndrome onset.

A PCOS diagnosis is challenging;[1] therefore, it is difficult to collect samples 'before' PCOS occurrence. Finding effective biomarkers for early PCOS screening may be beneficial. From our study, Mycoplasma colonization could be useful for PCOS screening, but specificity was insufficient, which might be because of the high prevalence of this opportunistic pathogen among healthy women. Future studies could focus on the pathogenic conditions of Mycoplasma and explore an accurate and convenient index for PCOS screening, such as Mycoplasma relative abundance. However, vaginal microbiota quantification faces serious technical challenges, such as swab collections where it is difficult to collect the same excretions from each participant.[43] Although our study was limited in not measuring the absolute vaginal microbiome, the relative microbiota abundance was practical and significant, as we had adequate sequencing depth and normalized the relative abundance corresponding to the least sequences when we analysed the sequencing data.

Other limitations of our study include lacking some important information (such as PH value for vaginal samples, use of vaginal lubricants, contraception methods), small sample size and some differences between groups at baseline (such as BMI), which might potentially impact the vaginal microbiome. Although we had adjusted these in the main analysis and included in subgroup analyses, the potential confounding effects exist. Meanwhile, multiple adjustments may make the sample size appear insufficient; the statistical power is limited. Further studies should match additional factors between groups at baseline, enlarge the sample size and collect more information to provide external validity and explore the internal mechanism.

In conclusion, there were distinct differences in vaginal bacteria between PCOS and healthy control women. The vaginal microbiome, especially the Mycoplasma element, has the potential to act as a biomarker in PCOS screening. Further research is required to evaluate the diagnostic value of the vaginal microbiome for PCOS and explore microbiota mechanisms underpinning this disease.