Impact of Gut Microbiota and SCFAs in the Pathogenesis of PCOS and the Effect of Metformin Therapy

Abstract

Polycystic ovary syndrome (PCOS) is a complex disorder that impacts both the endocrine and metabolic systems, often resulting in infertility, obesity, insulin resistance, and cardiovascular complications. The aim of this study is to investigate the role of intestinal flora and its metabolites, particularly short-chain fatty acids (SCFAs), in the development of PCOS, and to assess the effects of metformin therapy on these components. SCFA levels in fecal and blood samples from women with PCOS (n=69) and healthy controls (n=18) were analyzed using Gas Chromatography-Mass Spectrometry (GC/MS) for precise measurement. Fecal microbiota were quantitatively detected by real-time polymerase chain reaction (PCR). To assess the efficacy of six months of metformin treatment, changes in the microbiota and SCFAs in the PCOS group (n=69) were also evaluated. The results revealed that women with PCOS exhibited a significant reduction in beneficial bacteria (namely, the C. leptum group and Prevotella spp.) alongside a notable overgrowth of opportunistic microorganisms (C. perfringens, C. difficile, Staphylococcus spp., and Streptococcus spp.). An overproduction of acetic acid (AA, FC=0.47, p<0.05) and valeric acid (VA, FC=0.54, p<0.05) suggests a link between elevated SCFAs and the development of obesity and PCOS. Interestingly, AA in the bloodstream might offer a protective effect against PCOS by ameliorating key symptoms such as high body mass index (r=-0.33, p=0.02), insulin resistance (r=-0.39, p=0.02), and chronic inflammation. Although serum SCFA levels showed non-significant changes following metformin treatment (p>0.05), the normalization of AA in the gut underscores that metformin exerts a more pronounced effect locally within the gastrointestinal tract. Furthermore, the study identified the most effective model for predicting the success of metformin therapy, based on serum concentrations of butyric acid (BA) and VA, achieving a 91% accuracy rate, 100% sensitivity, and 80% specificity. These promising findings highlight the potential for developing targeted interventions and personalized treatments, ultimately improving clinical outcomes for women with PCOS.

Keywords: gas chromatography-mass spectrometry; intestinal microbiota; metabolomics; pathogenesis; polycystic ovary syndrome; short-chain fatty acids.

Figure 1

Correlation plot of body composition and clinical parameters in the control ($n=18$) and PCOS ($n=69$) groups before therapy. Cross sign marks statistically insignificant ($p>0.05$) relationship. The size of the dots associate with absolute value of correlation coefficient: stronger relationship—larger dots.

Figure 2

The influence of PCOS on gut microbiota before treatment, presented as a a radar chart (logarithmic scale) with median values of the number of colony-forming units: PCOS (red color, $n=69$) and control group (blue color, $n=18$). The data represent the average estimate of the ${log}_{10}$ of fecal real-time polymerase chain reaction (PCR) target genetic amplicon copy numbers in 1 g of feces. Statistically significant alterations ($p<0.05$) are indicated by an asterisk (*).

Figure 3

Violin plots for SCFA concentrations, relative abundances, and ratios (feces). Blue: the control group ($n=18$); red: the group with PCOS before treatment ($n=69$); yellow: the group with PCOS after treatment ($n=32$). AA: acetic acid; VA: valeric acid; BA: butyric acid; PA: propionic acid. * denotes $p<0.05$, ** denotes $p<0.01$, and *** denotes $p<0.001$, according to the Mann–Whitney test.

Figure 4

Clinically relevant parameters in the PCOS group ($n=38$) and their statistically significant associations with serum SCFAs before metformin therapy. Cross sign marks statistically insignificant ($p>0.05$) relationship. The size of the dots associate with absolute value of correlation coefficient: stronger relationship—larger dots.

Figure 5

PCOS group gut microbiota composition before treatment with expected effect from therapy. Red: radar chart for full effect of therapy; black: no effect of therapy. Statistically significant alterations ($p<0.05$) are indicated by an asterisk (*).

Figure 6

Receiver operating characteristic (ROC) curves, obtained by leave-one-out cross-validation of XGBoost models to predict the full effect of metformin therapy. Blue: model based on fecal SCFAs; red: based on fecal microbiota; green: combined data of fecal SCFAs and microbiota; black: based on serum SCFAs. AUC: area under the curve.