Exploring gut microbiota alterations and their associations with clinical symptoms in fibromyalgia Authors: Teresa Lopez de Coca, Francisco J Muñoz-Almaraz, Juan F Martinez-Blanch, Lucrecia Moreno Published: 2025 ABSTRACT AbstractFibromyalgia (FM) is a chronic pain syndrome characterized by widespread musculoskeletal pain, fatigue, sleep disturbances, and cognitive and gastrointestinal symptoms. It is classified as a central sensitization syndrome, involving altered pain processing in the absence of structural damage. Despite its high prevalence—affecting 2–4% of the global population, predominantly women—its pathophysiology remains poorly understood, and diagnosis is often delayed due to the lack of reliable biomarkers. Emerging evidence suggests that gut microbiota alterations may contribute to FM symptoms through immune dysregulation, neuroinflammation, and disruption of the gut-brain axis. Further studies controlling for these variables are needed to clarify the relationship. This study aimed to characterize the gut microbiota of women with FM, examine associated lifestyle and psychosocial factors, and compare them with healthy controls (HC). In this cross-sectional study, FM patients and HC underwent 16 S rRNA sequencing. Microbiota diversity were analyzed using QIIME2 and R. Thirty-seven FM patients and 46 HC participated in this study. FM patients reported greater pain (NRS-11: 5.92 vs. 1.05, p-value < 0.001), lower muscle mass (p-value: 0.021), and poorer nutrition (p-value: 0.021). They showed reduced α-diversity (Shannon index, p-value: 0.009) and distinct β-diversity (p-value < 0.001). SIBO prevalence was higher in FM (59.5% vs. 32.6%, p < 0.001), with methanogen dominance. FM patients had increased Akkermansia, Oscillospira, Methanobrevibacter, and Christensenellaceae. Network analysis revealed altered microbial interactions. Women with FM show less diverse gut microbiota and different microbial profiles. These findings suggest that the microbiota could represent a potential target for future diagnostic and therapeutic strategies, although further studies are needed to establish causal relationships.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-32101-y.Keywords: Bacterial overgrowth, Fibromyalgia, Microbiota, PainSubject terms: Pathogenesis, Rheumatology Introduction Fibromyalgia (FM) is a chronic pain syndrome characterized by widespread pain and hypersensitivity to stimuli. It belongs to the group of central sensitization syndromes, in which central pain processing mechanisms are altered. Unlike nociceptive pain, caused by tissue damage, or neuropathic pain, resulting from nerve injury, nociplastic pain (as observed in FM) occurs without structural damage to the somatosensory system1. The absence of specific biomarkers complicates diagnosis, delaying treatment and significantly impairing patients’ quality of life. FM affects 2–4% of the global population, with a female-to-male ratio of approximately 10:12. Beyond chronic pain, it is also associated with fatigue, stiffness, sleep disturbances, cognitive impairment, psychological disorders, and gastrointestinal dysfunction3. Emerging evidence suggests that gut microbiota alterations may contribute to FM pathophysiology, particularly through immune dysregulation and neuroinflammation. The gut microbiota, composed of trillions of microorganisms, plays a fundamental role in immune homeostasis and overall health. In FM patients, as well as in other pathologies, dysbiosis may enhance inflammation, increase intestinal permeability, and sensitize the nervous system to pain3,4. Moreover, alterations in bacterial metabolites—particularly short-chain fatty acids and lipopolysaccharides—have been observed in FM, suggesting a role in symptom modulation. Additionally, the gut microbiota can influence neurotransmitter pathways indirectly, particularly through the modulation of serotonin precursors, glutamate, and gamma-aminobutyric acid (GABA)5. Despite these findings, microbial alterations in FM remain differences across studies. While some reports finds an increase in Prevotella, Bacteroides, Phascolarctobacterium, Lachnoclostridium and Clostridium, or decreased levels in Faecalibacterium, Bifidobacterium and Bacteroides, others describe a decline in microbial diversity without identifying specific genera6–10. Similarly, in other rheumatic diseases, gut microbiota changes include increased Escherichia, Enterococcus, and Klebsiella, alongside a reduction in Bifidobacterium and Lactobacillus11. However, the limited number of studies and small cohort sizes, aligned with different dietary patterns, or confounding factors such as medication use, and pre-existing conditions may contribute to these discrepancies. To address this gap, this study aims to compare gut microbiota composition between FM patients and a well-defined healthy control group, rigorously excluding individuals with medication use or diagnosed pathologies. By characterizing the microbial profile in FM, this research could provide a foundation for future microbiota-targeted interventions, such as probiotic and dietary therapies, or fecal transplantation, to improve symptom management and patient outcomes. Recent research exploring the relationship between gut microbiota and FM has begun to uncover potential links between microbial imbalances and symptom severity, highlighting the importance of understanding the gut-brain axis and its implications for FM. Therefore, this exploratory analysis investigates associations between gut health, dysbiosis, and microbiota composition with FM symptoms and overall patient well-being. While causal mechanisms cannot be inferred from this design, identifying such relationships may help guide future longitudinal and interventional studies. Methods Study design This is a cross-sectional observational study designed to characterize gut microbiota differences between female FM patients and healthy controls (HC) women. This design allows for a snapshot analysis of microbial composition in both groups, identifying potential associations with FM. The woman pain memory study was reviewed and approved by the Institutional Review Board of Universidad CEU Cardenal Herrera (CEEI22/327, date of approval: 14 October 2022). The study was conducted in accordance with the ethical standards of the institutional and/or national research committee and with the Declaration of Helsinki and its later amendments and ethical standards. Participants Participants were recruited from Fibromyalgia Patient Associations and European Musculoskeletal Institute from Valencia (Spain), between September 2023 to March 2024. Inclusion criteria: Women aged ≥ 50 years. For the FM group were a clinical diagnosis based on the 2016 criteria of the American College of Rheumatology and reported widespread pain in at least 11 of 18 tender points. HC were selected based on the absence of diagnosed conditions and medication use. Nutritional status and BMI were assessed, ensuring no significant differences between groups. Exclusion criteria were applied to both groups and consist of women < 50 years or male participants, recent use (within the last three months) of antibiotics, probiotics, or prebiotics, or changes in their chronical treatments (e.g., antidepressants, anxiolytics, opioids or NSAIDs) or the presence of chronic gastrointestinal diseases such as irritable bowel syndrome, inflammatory bowel disease, or celiac disease. To ensure accurate identification of functional gastrointestinal disorders, participants were screened using a symptom-based questionnaire aligned with the Rome IV criteria, and those meeting diagnostic thresholds for IBS or other FGIDs were excluded. Additionally, the diagnosis of neurological or psychiatric disorders was previously established by specialized clinical services (e.g., neurology or psychiatry), and the final selection was made by rheumatologists who referred eligible participants for the study. The decision to include only women was based on the high prevalence of fibromyalgia in females (approximately 9:1 compared to males), aiming to avoid physiological variability related to sex. Participants under 50 years were excluded to reduce variability in cognitive outcomes, as cognitive decline can begin up to 20 years before clinical dementia diagnosis. Variables All participants underwent a comprehensive assessment, including cognitive, psychological, and physical evaluations. Cognitive function was evaluated using the Memory Impairment Screen12,13, the Short Portable Mental State Questionnaire, and the Semantic Verbal Fluency Test (SVF)14. Furthermore, depressive symptoms were evaluated using the Geriatric Depression Scale (GDS)15. Clinical pain intensity was evaluated using the Numeric Rating Scale (NRS-11)16. Finally, sleep disorders were evaluated using the STOP-BANG17 and Jenkins Sleep Scale18 questionnaires. Nutritional status was analyzed using the 14-item Mediterranean Diet Adherence Questionnaire19 and the Mini Nutritional Assessment (MNA)20. Anthropometric measurements were obtained using a Tanita InnerScan® bioimpedance analysis, assessing weight, BMI, body fat percentage, water percentage, muscle mass, metabolic age, and visceral fat. Additionally, circumferences of the arm, waist, and hips were measured with a tape measure. Small intestinal bacterial overgrowth (SIBO) testing SIBO detection was performed using Airtest kits, which quantified hydrogen and methane concentrations in exhaled breath following lactitol ingestion. Although lactitol is not among the most commonly used substrates, it was selected based on its availability and existing evidence supporting its diagnostic utility. These kits were generously donated by Calderón Laboratories (Castellon, Valencian Community, Spain). The test comprised eight breath sample collection tubes, with samples obtained over approximately three hours. Participants were instructed to fast for 12 h before testing and to refrain from using toothpaste or smoking. Patients reviewed the instruction manual beforehand and could contact the principal investigator for clarifications if necessary. Upon completion, the test kit, containing anonymized breath samples, was returned to the clinic of origin. Only the principal investigator and the patient’s physician had access to the results, with the physician determining further clinical management as appropriate. Microbiota DNA extraction and sequencing Fecal samples were collected using sterile containers with 5 mL cryopreservative (RNAlater™, Thermo Fisher Scientific) to maintain sample integrity. Samples were frozen at -80 °C immediately after collection. Genomic DNA from faecal samples (200 mg), 1 extraction blanks as negative controls, and 1 mock communities of microbes (ZymoBIOMICS™ Microbial Community Standard, Zymo Research Corporation, California, USA) were extracted using a combination of mechanical and enzymatic lysis, followed by QIAsymphony PowerFecal Pro DNA Kit (Qiagen) tracked the manufacturer’s protocol. Genomic DNA purity and yield were quantified with the NanoDrop Lite spectrophotometer (Thermo Fisher Scientific, USA). Genomic DNA samples were subjected to high-throughput sequencing following the 16 S Metagenomic Sequencing Library Illumina 15,044,223 B protocol (ILLUMINA)21. Briefly, v3-v5 region was PCR amplified using 16 S rRNA gene universal primers. In the second and last assay a universal linker sequence together with indexes and sequencing primers by Nextera XT Index kit (ILLUMINA) were included with PCR. 16 S based libraries were quantified by fluorimetry using Quant-iT™ PicoGreen™ dsDNA Assay Kit (Thermofisher). Libraries were pooled prior to sequencing on the MiSeq platform (Illumina), 300 cycles paired reads configuration. PhiX Control library (v3) (Illumina) was combined with the amplicon library (expected at 20%). The extraction and sequencing of the samples was carried out in Biopolis -S.L. (ADM R&D division, Paterna, Valencian Community, Spain). Paired-end FASTQ sequence files were imported into QIIME 2 (version 2024.2.0)22 and demultiplexed using the q2-demux plugin. Control quality was carried out with a minimum of 5.000 reads per sample, with an average sequence length exceeding 400 nucleotides were excluded. Reads were quality-filtered, primers were removed with q2-cutadapt, discarding untrimmed sequences. Reads were truncated at 250, merged, and denoised to Amplicon Sequence Variants (ASVs) using DADA2 through the q2-dada2 plugin23. The q2-rescript plugin24 was used to construct a reference-based phylogenetic tree of ASVs using the SILVA 138.1 database25. Taxonomy was assigned using the q2-feature-classifier plugin scikit-learn naïve Bayes classifier26 trained against the SILVA database trimmed to include only the v3-v5 region. QIIME 2 files were imported into R as a phyloseq object using qiime2R (https://github.com/jbisanz/qiime2R). Uncharacterized and unassigned phylum were removed. The raw data for this study have been deposited in the European Nucleotide Archive (ENA) at EMBL-EBI under accession number PRJEB89963. Statistical methods Microbial diversity was assessed using α-diversity and β-diversity metrics, implemented in the phyloseq and vegan R packages. α-diversity, which evaluates species richness and evenness within individual samples, was quantified using the Shannon and Chao1 indices. β-diversity, which examines differences in microbial composition between groups, was analyzed using Bray-Curtis dissimilarity and visualized through Principal Coordinates Analysis (PCoA). The statistical significance of β-diversity differences between groups was tested using permutational multivariate analysis of variance (PERMANOVA) with the adonis2 function from the vegan package. To identify potential confounding factors, a non-parametric Wilcoxon rank-sum test was performed on all the collected continuous covariates, and the rank-biserial correlation (rrb) was also calculated to estimate the effect size of group differences. Categorical variables were analyzed using Chi-Squared (χ2) and Fisher’s exact test. The Benjamini-Hochberg correction was applied to adjust for multiple comparisons, ensuring control of the false discovery rate (FDR). Additionally, Canonical Correspondence Analysis (CCA) and Correspondence Analysis (CA) were conducted using the vegan package to evaluate the influence of covariates on microbial composition (including all variables listed in the variables section). Principal Component Analysis (PCA) was applied to the covariate dataset to detect potential clustering patterns indicative of bias. For the differential abundance analysis, the ANCOM-BC (Analysis of Composition of Microbiomes with Bias Correction) method, implemented in the ANCOMBC package, was employed to adjust for compositional biases and account for false discovery rate (FDR). ANCOM-BC was parameterized to analyze only taxa with prevalences at least 20% less than prv_cut will be excluded in the analysis, aiming to reduce artifacts associated with sparsity. To investigate microbial interactions, a bacterial co-occurrence network was constructed using Spearman’s rank correlation coefficients, calculated with the Hmisc package. Correlations were considered significant at FDR-adjusted p-values < 0.05. The network structure was visualized using igraph and ggraph, applying a force-directed layout to highlight key associations among microbial taxa. All statistical analyses and visualizations were conducted in R (v.4.4.2). Ethics approval The woman pain memory study, was reviewed and approved by the Institutional Review Board of Universidad CEU Cardenal Herrera (CEEI22/327, date of approval: 14 October 2022). Consent to participate All participants provided written informed consent as established in the European General Data Protection Regulation (RGPD) and Organic Law 3/2018. Results Characteristics of the study population A total of 83 women participates, 37 patients with FM (mean age: 57.78 ± 4.85 years; mean BMI: 25.87 ± 4.88 kg/m²) and 46 HC (mean age: 57.71 ± 3.18 years; mean BMI: 26.24 ± 4.51 kg/m²) provided written informed consent and were included in the study. Notably, 75% of FM patients presented with gastrointestinal complaints, including constipation and abdominal bloating. The chronic medication consumed by the FM group is shown in Table 1. However, although participants in the FM group reported taking a wide variety of medications, these were generally administered at low doses, making it unlikely that medication use alone could fully account for the microbial differences observed in the study. Table 1Chronic treatments consumed by FM patients.FM TreatmentsATC codeProportionCorticosteroidsH02A5.40%NSAIDM01A27.03%Muscle relaxantM03B10.81%BisphosphonatesM05B13.51%OpioidsN02A13.51%AnalgesicsN02B47.06%AntiepilepticsN03A14.71%AntipsychoticsN05A8.11%AnxiolyticsN05B51.35%Hypnotics and sedativesN05C5.40%AntidepressantsN06A54.05% Chronic treatments consumed by FM patients. The FM group reported significantly higher pain scores (5.92 ± 1.69 vs. 1.05 ± 1.16 in HC, rrb: -0.98; adjusted p-value < 0.001), lower muscle mass (39.24 ± 4.12 kg vs. 42.03 ± 4.11 kg, rrb: 0.38; adjusted p-value: 0.032), and poorer nutritional status (12.16 ± 2.45 vs. 13.41 ± 0.91, rrb: 0.33; adjusted p-value: 0.041). Additionally, GDS (1.00 ± 1.48 vs. 0.00 ± 0.00, rrb: -0.55; adjusted p-value: < 0.001) and SVF (21.00 ± 8.90 vs. 27.00 ± 5.93, rrb: 0.46; adjusted p-value: 0.006) also showed statistically significant differences after adjustment (Table S1). When the association of these variables with microbiota composition at genus level was studied, only pain scores, muscle mass, and nutritional status were retained in the CCA (Figure S1), and were subsequently tested in the CA. Although neither the CCA nor the CA revealed a strong correlation between the evaluated covariates and the composition of the intestinal microbiota (Fig. 1A–C), a tendency towards dysbiosis was observed with the increase in NRS-11 values (Fig. 1A). The NRS-11 variable is the numerical variable that most effectively dissociates both categories, simply because of the fact of suffering from it. This is why this relationship with FM is considered significant. A–C Relationship between microbiota and significant co-variables with CA: NRS-11 (A), Muscle Mass (B) and MNA (C). D PCA of covariables. E PCoA of microbiota composition. F α-diversity by Shannon index. G α-diversity by Chao index. Additionally, the differences between the covariate sets in each group were assessed using PCA analysis. Subsequently, a PERMANOVA test using Euclidean distance did not presented statistically significant differences (Fig. 1D, p-value: 0.375). No significant differences (adjusted p-value > 0.05) were observed between the groups in adherence to the Mediterranean diet, sleep quality, or anthropometric variables, including body fat percentage, water percentage, metabolic age, and waist-to-hip ratio. Once the absence of bias in the covariates of both groups was confirmed, we aimed to investigate the differences in microbial composition. Microbial diversity Significant differences were found in β-diversity with Bray-Curtis metrics (Fig. 1E, p-value < 0.001) and α-diversity for Shannon index for ASV, with lower α-diversity observed in the FM group (mean diversity in FM: 4.15, mean diversity in HC: 4.34, Fig. 1F, p-value: 0.009), however, Chao index didn’t show significance (Fig. 1G, p-value: 0.701). Small intestinal bacterial overgrowth study The presence of SIBO with hydrogen or methane overproduction was assessed. As shown in Table 2, a large proportion of FM patients were classified in the methanogenic SIBO group (59.46%), while the HC group exhibited a considerably lower prevalence (32.61%). This revealed a significant association between the condition and SIBO presence (p-value < 0.001). Table 2Small intestinal bacteria overgrow distribution.NegativeOvergrow H2Overgrow CH4p-value Fisher ‘s Exact TestFM8 (21.62%)8 (21.62%)21 (56.76%) < 0.001 *** HC30 (65.22%)3 (6.52%)13 (28.26%)TOTAL381134 Small intestinal bacteria overgrow distribution. Additionally, all bacteria involved in hydrogen production, methanogenic, sulfate-reducing, and acetogenic activity were analyzed together to assess their potential confounding effect on microbial composition differences (see Table S2 for the list of genera included in each group27–39. Significant differences were found in PERMANOVA using Bray-Curtis distance between FM and HC (p-value: 0.031). These differences remained significant even after excluding all hydrogen-producing bacteria (p-value: 0.023). However, although a trend was observed, no clear distinction emerged based on the type of SIBO detected. An analysis of beta-diversity was carried out selecting only bacteria involved in hydrogen production, methanogenic, sulfate-reducing, and acetogenic, (Supplementary Materials). Significant differences were found by means of PERMANOVA using Bray-Curtis distance between FM and HC (p-value: 0.031). These differences remained significant even after excluding all hydrogen-producing bacteria (p-value: 0.023). No significant differences in beta-diversity were found depending on the type of SIBO (p-value: 0.302). Differential abundance analysis After analyzing the covariates and ruling out their association with microbial composition, as well as confirming significant differences in alpha and beta diversities between the groups, we proceeded with a differential abundance analysis using ANCOM-BC. Table 3 highlights the findings. In the FM group, there was a significant increase in the genera Akkermansia, Oscillospira, Methanobrevibacter, Erysipelatoclostridium, and an uncultured genus from the Christensenellaceae family, after p-value adjustment. Table 3Differential abundance of bacterial genera identified by ANCOM-BC analysis.Log Fold Changep-valueAdjust p-value Akkermansia 1.84450030.0000000 0.0000012 *** Oscillospira 0.90350100.0000087 0.0009714 *** Methanobrevibacter 1.06614280.0000111 0.0012269 ** Erysipelatoclostridium 0.95155230.0000799 0.0087868 ** Christensenellaceae_uncultured 0.84219470.0003354 0.0365628 * Anaerotruncus 0.79449820.00052380.0565715 Clostridia_UCG-014 -1.16921300.00083070.0888803Log Fold Change represents the variation in FM group vs. HC. Significant p-values are indicated in bold. *p-value < 0.05; *: p-value < 0.01; **: p-value < 0.001; **. Differential abundance of bacterial genera identified by ANCOM-BC analysis. Log Fold Change represents the variation in FM group vs. HC. Significant p-values are indicated in bold. *p-value < 0.05; *: p-value < 0.01; **: p-value < 0.001; **. Although the genus Anaerotruncus exhibited a similar upward trend, with a false discovery rate lower than 0.1. Conversely, the genus Clostridia_UCG-014 displayed a decreasing trend in the FM group, but this difference was not statistically significant. The ANCOM-BC sensitivity analysis indicates that the differences for the genus Anaerotruncus, in addition to Oscillospira, were robust. Figure 2A provides a graphical representation of the log fold change (lfc) for the taxa analyzed. In the ANCOM-BC heatmap, positive lfc values indicate genera enriched in the FM group, while negative values reflect enrichment in the control group. The color intensity corresponds to the magnitude of the lfc, with brighter colors highlighting larger differences between groups. This visualization complements the statistical findings by emphasizing key taxonomic shifts in the FM group. A Heatmap of significantly different bacterial genera. B Bacterial co-occurrence networks in individuals with FM. C Bacterial co-occurrence networks in HC. Green lines represent positive correlation. Red lines represent a negative correlation. After analyzing Spearman’s correlations, two distinct patterns of microbial modulation emerged. First, the genus Clostridia_UCG-014 demonstrated extensive correlations with other bacterial genera in the HC group, but these correlations were entirely absent in the FM group (Fig. 2B). Conversely, the genus Prevotella_9 and Faecalibacterium acted as a primary modulator of bacterial correlations within the FM group, despite showing a lower relative abundance than HC, where no such correlations were observed (Fig. 2C). Additionally, in the HC group, the genus Dialister showed no significant correlation with the genus Christensenellaceae. Discussion This study highlights significant differences in gut microbiota profiles between FM patients and HC. These findings underscore the potential role of microbial dysbiosis in FM pathophysiology. Adjustments for covariates like age, sex, anthropometric variables, and lifestyle factors did not alter these microbiota differences, supporting the robustness of the observed associations. A notable result was the positive correlation between SIBO and FM, suggesting a potentially relevant association between microbial imbalances and the disease. Our analysis revealed reduced α-diversity in FM patients compared to HC, alongside significant β-diversity differences. These results consistent with prior studies6–9,40 that establish microbial diversity as a hallmark of dysbiosis in FM. Reduced microbial diversity likely reflects a disrupted gut ecosystem that may confer to FM symptoms. A correlation network analysis further revealed alterations in specific taxa and their interactions within the FM group, highlighting microbial imbalances that may contribute to the systemic and central symptoms associated with FM. Differential abundance analysis (ANCOM-BC) identified significant enrichment of Akkermansia, Oscillospira, Methanobrevibacter, Erysipelatoclostridium, and an uncultured genus from Christensenellaceae in FM patients. Akkermansia, while beneficial in producing short-chain fatty acids (SCFAs), may exacerbate mucosal erosion when overrepresented, contributing to increased intestinal permeability and systemic inflammation. Similarly, Oscillospira, linked to eubiosis and gut health, is elevated in FM and in our population, has also been associated withconstipation, a common symptom in FM41. Methanobrevibacter and Christensenellaceae, bacteria with strong positive correlations, are enriched in FM and are associated with SCFA production and reduced inflammation under normal conditions. Their overrepresentation in FM may reflect compensatory responses to dysbiosis. Notably, increased Methanobrevibacter correlates with SIBO and methane overproduction, conditions prevalent in FM (56.76%). Additionally, Erysipelatoclostridium, elevated in FM, is an opportunistic pathogen potentially contributing to disease pathology. Furthermore, microbiota studies have shown that bacterial taxa predominantly represented by Oscillospira, Ruminococcaceae, RF39, Rikenellaceae, Akkermansia, and Christensenellaceae are positively correlated among them42. In our patients, however, Ruminococcaceae, Rikenellaceae, and RF39 exhibited reduced abundance compared to the increased levels of Oscillospira, Akkermansia, and Christensenellaceae. This highlights distinct correlation networks in FM and suggests dysbiosis as a potential biomarker for the condition. On the other hand, although not significantly, our findings align with previous studies that evaluated ß-diversity6–9, which report an increase in Phascolarctobacterium and Lachnoclostridium in FM patients, along with a decrease in Faecalibacterium, Bifidobacterium and Bacteroides. Variations in the statistical significance of specific bacterial taxa across published studies may stem from variations in the differential abundance analysis methods employed. While DESeq2 is commonly applied in microbiome research, it tends to generate a higher rate of false positives, particularly when data are filtered to retain only the most abundant taxa. In contrast, ANCOM-BC offers greater control over false discoveries and produces more reliable estimates in the context of compositional data43. For this reason, ANCOM-BC was selected for our analysis. These methodological differences may partly account for the discrepancies observed between our results and those of previous studies. Spearman correlation analysis revealed distinct microbial association patterns in FM. For example, Clostridia, a key producer of butyrate, is reduced in FM and shows strong correlations in HC (Fig. 2C). Butyrate supports gut health by enhancing barrier function and reducing inflammation, and its depletion may exacerbate systemic symptoms in FM. In contrast, Prevotella and Faecalibacterium, although less abundant, show strong correlations in FM group (Fig. 2B), that are not present in HC, indicating unique interaction patterns. The high prevalence of intestinal dysbiosis and reduced Prevotella abundance in our FM population may have relevant pathophysiological consequences. Prevotella plays a key role in immune modulation via TLR2 signaling and is negatively associated with inflammation. Its reduction may contribute to immune dysfunction and chronic inflammation44,45. As a major contributor to microbial diversity and fiber metabolism, lower Prevotella levels may also impair gut resilience and short-chain fatty acid production46. These changes could help explain the high rates of comorbid conditions in FM, including rheumatoid arthritis, mood disorders, and irritable bowel syndrome46–48. Specifically, Prevotella spp. have been identified as key pathobionts in rheumatoid arthritis, with significant enrichment observed in patient fecal samples49,50. A 27 kDa protein-derived peptide produced by Prevotella can mediate Th1 cell immune responses51,52. Furthermore, gut microbes and their metabolites exert a dual regulatory role, highlighting their importance in modulating immune processes. Overall, our findings support a role for gut microbial alterations in FM pathophysiology. Finally, the reduction in Faecalibacterium, a known anti-inflammatory bacterium, further underscores the dysbiosis observed in FM. This reduction parallels findings in inflammatory bowel disease but contrasts with the inverse relationship observed in Akkermansia, suggesting a distinct microbial signature for FM. The altered abundance and correlation patterns of these taxa may serve as biomarkers for FM and offer new therapeutic targets. Furthermore, the observed microbial dysbiosis may have implications beyond the gut. The gut-brain axis is a bidirectional communication network linking the gut microbiota to the central nervous system. Emerging evidence suggests that microbial composition influences neurotransmitter balance, particularly GABA, glutamate, and serotonin53–68. Certain bacteria can produce or modulate these neurotransmitters, affecting brain function and behavior. Dysbiosis may disrupt this balance, potentially contributing to the pathophysiology of disease. Previous studies have shown reduced GABA levels in specific brain regions of FM patients53,54. This deficiency is linked to impaired inhibitory neurotransmission and increased insomnia and pain sensitivity53,69. In response, an upregulation of GABA receptors has been observed in multiple brain areas70. This suggests an attempt to compensate for GABAergic dysfunction, which correlates with altered pain processing and network activity. Additionally, lower GABA levels have been associated with reduced pressure-pain thresholds, as well as increased sleep problems, further supporting the role of impaired inhibitory control in FM pathophysiology54. Emerging evidence suggests that gut microbiota may influence GABA production53–56. Our findings align with reports linking reduced GABA levels to lower Lactococcus and Enterococcus abundance54,55. However, unlike other studies53,56, we observed an increase in Lactobacillus and Escherichia. These discrepancies may reflect underlying FM pathophysiology, emphasizing the need for further research into the microbiota-GABAergic system interaction. Central sensitization also plays a key role in FM. This condition heightens pain sensitivity through NMDA receptor overactivity, driven by increased glutamate levels. Elevated glutamate concentrations have been reported in key brain regions, such as the insula, hippocampus, and posterior cingulate cortex66. These changes correlate with enhanced pain perception, further implicating excitatory-inhibitory imbalance in FM pathology. The interplay between glutamate and GABA appears crucial. Excessive glutamate, combined with deficient GABA, may contribute to FM symptoms by amplifying pain transmission and weakening pain inhibition67. Interestingly, Lactobacillus, which was increased in our FM cohort, can produce significant amounts of glutamate under specific metabolic conditions68. Moreover, glutamic acid decarboxylase (GAD) plays a crucial role in maintaining the glutamate-GABA balance by catalyzing the conversion of glutamate to GABA57. Reduced GAD expression or activity, particularly of the GAD65 isoform, has been reported in FM, potentially exacerbating neurotransmitter dysregulation. While previous studies suggest that stress, diet, obesity, sedentary behavior58, and gut microbiota59–61 influence GAD activity, our patients did not differ from controls in co-variables. This finding underscores a potential role of the gut microbiota in FM pathophysiology, independent of commonly analyzed factors. Our findings align with previous studies suggesting that GAD levels may decrease when the relative abundance of Bifidobacterium60 and Escherichia61 is reduced, whereas an inverse relationship is observed for Lactobacillus59. On the other hand, serotonergic dysfunction plays a key role in the pathophysiology of FM affecting both ascending and descending pain pathways. Patients with FM exhibit significantly lower serum serotonin levels compared to HC71. This reduction is associated with increased symptom severity, including pain, anxiety, and depression62,71. In particular, descending serotonergic fibers at the spinal level are crucial for modulating nociceptive processing, which regulates pain perception62. Serotonin production in the gut is significantly influenced by microbiota. Ruminococcus, which was found to be reduced in our FM cohort showing a less convenient microbiome, is known to stimulate peripheral serotonin production. This bacterium metabolizes dietary phenylalanine and tryptophan into phenethylamine and tryptamine, which, in turn, promote serotonin and melatonin biosynthesis in enterochromaffin cells. Specifically, R. gnavus enhances serotonin production through the activation of the trace amine-associated receptor 164. Serotonin, primarily synthesized in the gastrointestinal tract, is secreted into the lumen and subsequently cleared by the serotonin transporter. Furthermore, serotonin has been shown to reduce the expression of virulence genes in Enterohemorrhagic Escherichia coli and Citrobacter rodentium, suggesting a potential role in modulating intestinal colonization by pathogenic bacteria65. Strengths and limitations This study has several strengths that support the validity and relevance of the findings. Notably, all data were collected through in-person, structured 90-minute interviews conducted by trained researchers, allowing for a detailed and reliable characterization of each participant. This protocol ensured accurate assessment of medication use, psychological profiles, nutritional intake (including both macro- and micronutrients), and clinical variables such as small intestinal bacterial overgrowth (SIBO). To our knowledge, few studies in this field have gathered such a comprehensive dataset under controlled and standardized conditions. However, the study also has limitations. The sample size was relatively small (n = 83), limiting generalizability and statistical power. The cross-sectional design prevents causal inference; therefore, microbial differences observed in fibromyalgia (FM) may reflect disease consequences, such as altered gut motility or immune dysregulation, rather than causal factors. The detection of SIBO used lactitol as a substrate, which, though validated, may limit comparability with studies using glucose or lactulose. Additionally, although gastrointestinal disorders were part of the exclusion criteria, no standardized symptom scale was applied. Given that 75% of FM patients reported GI complaints, this limits our ability to explore associations between microbial patterns and specific GI phenotypes. Conclusions This study identifies significant gut dysbiosis in FM patients. Reduced α-diversity and distinct β-diversity patterns were observed compared to HC. Altered microbial composition included a depletion of beneficial taxa, alongside an overrepresentation of Akkermansia, Oscillospira, and Erysipelatoclostridium. These changes may contribute to immune dysfunction, chronic inflammation, and systemic symptoms in FM. These findings suggest that gut microbiota alterations play a critical role in FM through the gut-brain axis. Dysbiosis may impair neurotransmitter balance, reduce GABA and serotonin levels while increase glutamate activity. These imbalances are linked to heightened pain sensitivity, sleep disturbances, and neuropsychiatric symptoms, underscoring the importance of microbiota in FM pathophysiology. The consistent overrepresentation of Akkermansia, Oscillospira, and Erysipelatoclostridium in FM patients suggests that these taxa may serve as potential biomarkers for diagnosis. Altered abundance and correlation patterns among bacterial taxa suggest gut dysbiosis as a central feature of FM. Although further research is needed, therapeutic strategies targeting microbial diversity, intestinal barrier integrity, and neurotransmitter production may offer promising complementary avenues for symptom management. These interventions should be integrated within a multidisciplinary framework, in line with current treatment guidelines combining pharmacological and non-pharmacological approaches. Supplementary Information Below is the link to the electronic supplementary material. Supplementary Material 1 Acknowledgements We would like to acknowledge all the patients that participated in this study for their collaboration. We also acknowledge Cathedra DeCo MICOF UCH University who funded this research. Abbreviations Analysis of Composition of Microbiomes with Bias Correction Amplicon Sequence Variants Correspondence Analysis Canonical Correspondence Analysis False Discovery Rate Fibromyalgia Glutamic aciddecarboxylase Geriatric Depression Scale Healthy controls Log fold change Mini Nutritional Assessment Numeric Rating Scale Principal Component Analysis Principal Coordinates Analysis Semantic Verbal Fluency Author contributions All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by Teresa Lopez de Coca, and Francisco J. Muñoz-Almaraz. The first draft of the manuscript was written by Teresa Lopez de Coca and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript. Funding This research was funded by the Cathedra DeCo MICOF UCH University. Teresa Lopez de Coca was supported by the Research Fellowship grant from “Ayudas a la Formación de Jóvenes Investigadores CEU-Santander”. Data availability The datasets used and/or analyzed during the current study are available from the corresponding author. All the analysis scripts and code used in this study are available at the following repository: https://github.com/fjmalmaraz/CBFM. Declarations Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary Material 1 Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author. All the analysis scripts and code used in this study are available at the following repository: https://github.com/fjmalmaraz/CBFM.