Gut Microbiome Variation Across Menstrual Phases: Distinct Roles of Diet and Physiological Hormonal Fluctuations in Healthy Women

Gut Microbiome Variation Across Menstrual Phases: Distinct Roles of Diet and Physiological Hormonal Fluctuations in Healthy Women | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Gut Microbiome Variation Across Menstrual Phases: Distinct Roles of Diet and Physiological Hormonal Fluctuations in Healthy Women Serena Sanna, Valeria Lo Faro, Paola Forabosco, Francesca Crobu, and 20 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9888793/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Sex differences in the gut microbiome are well established, yet the contribution of reproductive hormone variations within women remains insufficiently characterized. Here, we investigated how hormonal fluctuations during a natural menstrual cycle shape gut microbiome in healthy women and whether these effects are distinct from dietary influences. Leveraging longitudinal shot-gun gut microbiome data from 160 Italian participants in the Women4Health cohort, combined with detailed dietary information, we examined relationships with five reproductive hormones hormones: 17β-estradiol, progesterone, follicle-stimulating hormone (FSH), luteinizing hormone (LH) and prolactin. Microbial alpha diversity was significantly lower at the follicular phase compared to ovulatory phase (p = 0.03), while beta diversity significantly increased from follicular to the late luteal phase (p = 8.7x10 − 6 ). Twenty-eight species and twenty-two microbial pathways were significantly associated with reproductive hormones, with the majority of associations observed for gonadotropins (FSH and LH) rather than gonadal steroids (17β-estradiol and progesterone). While we confirm strong relationships between diet and microbiome, mediation analysis showed no evidence for a mediating effect of hormones on those. Together, these findings reveal distinct hormonal and dietary influences on gut microbiome across the menstrual cycle in healthy women. Biological sciences/Computational biology and bioinformatics/Data processing Health sciences/Medical research/Epidemiology Biological sciences/Microbiology/Microbial genetics/Bacterial genes Health sciences/Risk factors Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Diet is a dominant factor associated with interindividual variability in gut microbiome. Recent large-scale studies based on well-characterized cohorts have documented how dietary patterns shape microbiome composition and, in turn, associated the microbiome with metabolic biomarkers indicative of cardiometabolic health 1 – 5 . For instance, Fackelmann and colleagues demonstrated that common dietary profiles (omnivore, vegetarian, and vegan) produce distinct gut microbiota signatures reflecting the predominant foods consumed in each diet. A diet with red meat consumption, for example, is marked by the presence of characteristic gut bacteria, such as Ruminococcus torques and Alistipes putredinis , which are largely absent in vegan individuals 5 . Notably, even single foods can leave unique microbial “fingerprints”, as in the case of coffee consumption, which is strongly associated with the presence of Lawsonibacter asaccharolyticus 6 . While research on the diet-microbiome interaction has advanced considerably, the influence of other host factors remains yet to be understood. For example, several studies have shown that sex is a significant determinant of microbiota diversity in human populations 7 , 8 , a difference largely attributed to sex steroid hormones. Men and women in fact diverge in gut microbiome composition—such as altered Firmicutes/Bacteroidetes ratios—and these differences emerge through puberty (Kim 2020; Sisk-Hackworth 2023; Özkurt 2021) and are reduced after menopause 9 – 11 . Moreover, gut bacteria are known metabolizers of steroid hormones. Several taxa express the beta-glucuronidase and beta-glucosidase enzymes, which hydrolyze conjugated (inactive) estrogens that have been excreted by the liver into the gut 12 . This process liberates active estrogens, enabling their reabsorption through enterohepatic circulation 13 – 15 . The collection of these microbes and their metabolic products capable of metabolizing estrogens have been termed the “estrobolome”. Both reduced gut microbial diversity and diminished microbial β‑glucuronidase activity observed in post-menopausal women, compared to pre-menopausal women, have been associated with systemic inflammation and cardiometabolic risk 16 . Other complex mechanisms involving microbial enzymes have been proposed for the metabolization of other steroid hormones, such as progesterone, but current knowledge remains limited 15 . Sex steroid hormones also play direct roles in digestive physiology. For example, estrogen contributes to gastrointestinal motility and mucosal health, and its decline during menopause has been associated with increased gastrointestinal symptoms 17 . Evidence from animal models further highlights the relevance of hormone-microbiome interactions: ovariectomized mice fed with a low-fat diet exhibit marked weight gain and increased inflammatory status, compared with high-fat-fed mice, suggesting that the metabolic impact of gut microbiome-diet interaction is contingent on female sex hormone status 18 . To date, studies in human cohorts have typically examined either sex differences or dietary influences, but few have considered diet and hormones simultaneously. Studies focusing on women have been predominantly cross-sectional, frequently contrasting healthy and pathological states 16 or comparing pre- and post-menopausal women 19 , 20 . Furthermore, only small-scale studies have directly measured reproductive hormones along with gut microbiome 21 , 22 . Conversely, many longitudinal studies on diet and microbiota lack direct information on reproductive hormones hormones and hormonal status or rely solely on the information on the number of days since menstruation 23 . To address these gaps, we propose a comprehensive study of gut microbiome profiles across the menstrual cycle in healthy Italian reproductive-age women from the Women4Health cohort 24 , incorporating detailed dietary assessments and direct measurements of reproductive hormones – including not only estrogen and progesterone, but also hormones not yet explored for their interaction with gut microbiota, such as follicle-stimulating hormone (FSH), luteinizing hormone (LH), and prolactin (PRL) (Fig. 1 ). By disentangling the relative contributions of diet and endogenous hormonal fluctuations to microbiome variation, we elucidate mechanisms uniquely relevant to female microbial ecology and health—with important implications for precision microbiome-based interventions in women. The figure describes the sampling scheme along the menstrual cycle and the data collected and analyzed in this work. Methods Study cohort description The Women4Health (W4H) study cohort is a longitudinal, population-based study designed to investigate the interplay between hormonal fluctuations, the microbiome, and metabolic health in women of reproductive age. The study is the result of a collaborative effort between the Institute of Genetics and Biomedical Research (IRGB) of the National Research Council (CNR), the IRCCS Burlo-Garofolo in Trieste, the Azienda Ospedaliero Universitario (AOU) of Cagliari, and the IRCCS Policlinico Sant’Orsola in Bologna, Italy. Recruitment is ongoing, with the aim of enrolling up to 300 healthy women aged 18–45 years with regular menstrual cycles and no major metabolic, gastrointestinal, or gynecological conditions. Full details on the study design, recruitment protocol, and inclusion criteria have been previously published 24 . Briefly, at enrollment and during four follow-up visits corresponding to specific phases of the menstrual cycle, detailed questionnaire data and multiple biological samples (stool, saliva, blood and vaginal swabs) were collected. W4H data collection took place in three Italian cities (Trieste, Cagliari and Bologna) representing the North-East, South, and Central-North regions of the country, respectively. The collection began at different times across these centers: in Trieste, it started in February 2022, in Cagliari in 2024 and in Bologna in 2025. At enrollment, volunteers enrolled were asked to complete a questionnaire. This captured a wide range of information, including demographic and anthropometric data, a detailed medical and gynecological history, family history of disease, exposure to tobacco (direct or indirect), use of prebiotics and probiotics and dietary habits through a food frequency questionnaire (FFQ). During follow-up visits, scheduled to capture four phases of the menstrual cycle (follicular (F), ovulatory (O), early and late luteal phases (EL and LL)), they completed additional questionnaires that included, among many items, information on stool consistency (using the Bristol Stool Form Scale) and the use of any medications or supplements. In addition to the food frequency questionnaire (FFQ), participants also recorded their daily dietary intake using MyFitnessPal, a user-friendly smartphone application that offers an extensive food database. The app allows for accurate tracking of daily calorie intake as well as macronutrients and other dietary components, including protein, fat, cholesterol, sugar, and fiber. Daily diaries were submitted by volunteers at the end of the observation period. From February 2023 to 1 August 2025, 161 participants were enrolled and returned for at least one visit. Among these, 98 were from Trieste, 38 from Bologna, and 25 from Cagliari. While the FFQ was available for all of them, only a subset of 149 compiled and returned the MyFitnessPal diary. Descriptive statistics of the 161 women are given in Supplementary Table 1. Measurement of hormone levels and reclassification of menstrual phases Blood samples were collected via venipuncture using silica gel-coated Vacutainer tubes. After collection, blood was left undisturbed at room temperature for 1 hour to allow clotting. Subsequently, serum was obtained by centrifugation at 4,000 rpm for 5 minutes at room temperature. A volume of 0.5 mL of serum was then aliquoted into two sterile safe-lock polypropylene tubes and stored at − 20°C. An aliquot of 1.5 ml serum was frozen and later used to measure 5 reproductive hormones: 17β-estradiol (E2), luteinizing hormone (LH), follicle-stimulating hormone (FSH), progesterone (P4) and prolactin (PRL). During each woman’s first visit, we also measured testosterone and thyroid parameters (thyroid stimulating hormone and free thyroxine (free-T 4 )) to exclude endocrine disorders. Hormones were measured in three separate batches (one for each year of recruitment: 2023, 2024, 2025) on the ADVIA Centaur CP Immunoassay System instrument (Siemens Healthineers). Hormone data processing involved three steps. First, we log-transformed the hormone measurements. Second, we corrected hormone levels for the batch effect by matching the means of batch 1 and 2 hormone distributions with the reference batch 3. For each hormone, we calculated the difference in means between batch 1 or 2 with batch 3 and subtracted this difference from the values of batch 1 and 2. Third, we found that serum storage time in the freezer affected hormone levels. To account for this, we fitted a Linear Mixed Model (LMM) with a fixed effect for storage time and a random intercept for ID (subject-specific baseline shift): hormone ~ storage_time + (1|ID), using lme4 R package v. 1.1–36 and took the residuals of this model by subtracting the predicted population average of the storage time effect on hormone levels from the real hormone level: pheno_real - predict(lmm_fit, re.form = NA) By study design, samples were collected at four visits scheduled according to the last day of menstruation and intended to capture the major hormonal phases of the menstrual cycle: follicular, ovulatory, early luteal, and late luteal. Since each woman exhibits a unique temporal pattern of reproductive hormones, days since last menstruation may not always represent an accurate proxy for menstrual phases. Therefore, we developed an algorithm to redefine phases in a more biologically meaningful way, using directly measured concentrations of LH. Details have been described in a previous work from our group 25 . Briefly, the procedure first determines the visit with maximal LH concentration as the ovulatory point. Once the LH peak is located, a predefined mapping shifts the remaining visits backward or forward while preserving their temporal order relative to ovulation. This shift may map some visits outside the valid phase range (1–4). Such boundary cases are resolved by assigning the visit to the closest phase. The final output provides, for every visit of every participant, a biologically informed phase label. When multiple visits map to the same phase, downstream analyses use the average value of all corresponding measurements (reproductive hormones, phenotypes, dietary parameters, or microbiome relative abundances). For some women a phase could not be established and thus they were excluded from analyses. While we recognize that the optimal way to establish phases would have been to have information on the day of the subsequent menstruation or echographic measurements, these were not available to us. Therefore, in this work we present comparisons of diet and microbiome data across menstrual phases being aware that these are not clinically evaluated but still provide the most robust approximation using directly measured concentrations of oestradiol, progesterone, FSH, and LH. Fecal sample processing In each center and at enrollment, volunteers received stool kits along with detailed instructions for self-collection. To ensure sample stability at room temperature, the OMNIgene-GUT (OM-200, DNA Genotek) collection system was used. After sample collection, volunteers returned their stool samples to the Obstetrics and Gynecology Clinic within 24 hours, where samples were stored at room temperature. After homogenization with a vortex mixer, 250 microL of stool were aliquoted in two 2.0 mL safe-lock polypropylene tubes and stored at -80°C Celsius. Microbial DNA was extracted using the PowerFecalPro DNA kit (QIAGEN) and subsequently stored at -30°C. A total of 551 samples were collected from 161 volunteers, across four visits. Microbial DNA extracted from these samples were shipped in two separate batches at Prebiomics S.r.l. and sequenced using the NovaSeq 6000 System platform (Illumina), targeting a sequencing output of approximately 9 Gb of 150 bp paired-end reads per sample. The first batch included samples from the initial 75 participants from Trieste, totaling 264 samples across the four visits. A second batch followed, repeating 1 sample that failed in batch 1 and completing the dataset with 287 samples. Profiling microbiome composition and function The metagenomes were first analyzed using the KneadData tool (v0.12.0) to process metagenomic reads in FASTQ format. The reads were trimmed to a PHRED quality score of 30, and Illumina adapters were removed. After trimming, the integrated Bowtie2 tool (v2.5.3) 26 within KneadData was employed to eliminate reads that aligned to the human genome. The total number of sequencing reads, remaining after quality filtering and removal of human-origin reads and DNA concentration calculated at the sequencing center, was used as a covariate, as this could influence the detection of specific microbial taxa. The taxonomic composition of 551 metagenomes from 161 volunteers across the three collection sites (Trieste = 98, Bologna = 38, Cagliari = 25) was profiled using MetaPhlAn (v4.1.1) with the MetaPhlAn marker gene database (mpa_vJun23) 27 . MetaPhlAn provided species-level abundance estimates, referencing the CHOCOPhlAn (mpa_CHOCOPhlAn_202403) database. Additionally, microbial biochemical pathways were profiled using the HUMAnN pipeline (v3.9) 28 , integrated with the DIAMOND alignment tool (v2.1.9), the UniRef90 protein database and the ChocoPhlAn pan-genome database. The analysis identified 12 phyla, 303 classes, 317 orders, 353 families, 1,043 genera, 1,740 species and 185 pathways. After quality control applied to metagenomes and assessment of metadata, 9 samples were excluded from subsequent analysis due to low microbial diversity, inconsistencies in reported menstrual cycle information, or discrepancies in cycle phase classification. Following these exclusions, 542 fecal samples from 160 women were retained (Trieste = 98, Bologna = 37, Cagliari = 25). When applying the redefined menstrual cycle phase criteria (see above), 489 samples from 160 women were included in the analyses. Processing of dietary information Dietary intake was assessed using two complementary approaches: a food frequency questionnaire (FFQ) and daily dietary records. The FFQ was administered at enrollment and was designed to estimate habitual intake of selected food items over typical daily or weekly time frames. All 161 participants with quality controlled (QCed) microbiome data provided the FFQ at enrollment, while daily diet registers were completed by a subset of only 149 participants. Specifically, they recorded all foods and beverages consumed throughout the study period using the MyFitnessPal application, which allows detailed documentation of food items and portion sizes and automatically calculates total energy (kcal) and nutrient intakes, including carbohydrates (g), fats (g), proteins (g), sugars (g), fiber (g), cholesterol (mg), and sodium (mg). Daily diet records were carefully manually screened for potential errors by identifying foods with outliers in nutrient intakes. Detailed information on dietary data processing and quality control is given in the Supplementary Notes . To evaluate the agreement between the two dietary assessment methods, daily total energy (kcal) and nutrient intakes were estimated from the FFQ-derived consumption frequencies (see Supplementary Notes ) and compared with the daily averages (across the entire observation period) obtained from the daily dietary records. Differences in mean intakes between the two methods were assessed using paired Wilcoxon signed-rank tests, while relative agreement was evaluated with Spearman rank correlation coefficients. To account for multiple comparisons, p-values were adjusted using the Bonferroni correction. All analyses were performed using custom R script available in our GitHub folder (see Data Availability Statement ). Differences in nutritional intake across the menstrual phases Longitudinal changes in nutrient intake along the menstrual cycle were analyzed using Linear Mixed Models (LMM) implemented with the lmer function from the lme4 package (v1.1-36) with p-values calculated via lmerTest package (v3.2-1). For each nutrient, we considered the 3-days average prior to fecal sample collection as variables (nutrient 3days ). Temporal variation was modeled using hormonal phase (treated as numeric, continuous variable) as a fixed effect, and participant ID was included as a random intercept to account for repeated measures. Age, BMI, and recruiting centers were included as covariates to control individual- and center-level variability. Thus, for each nutrient we fit the following model: nutrient 3days ~ phase + age + BMI + center + (1|ID) Multiple testing across nutrients was corrected for using the Bonferroni method. Potential non-linear effects of time were assessed initially using natural splines implemented in the splines package (v4.4.3). Spline models were compared with reduced models via likelihood ratio tests, and, if significant, a linear model was also fitted and compared with the spline model to determine whether the relationship was essentially linear or non-linear. Predicted marginal effects of the main temporal variable were obtained using the effects package in R (v4.2-4) and visualized as line plots to illustrate nutrient trajectories across hormonal phases. In addition to LMM, pairwise comparisons of nutrient intake between hormonal phases were conducted using paired Wilcoxon signed-rank tests to account for repeated measures within individuals. Analyses were performed on age- and BMI-adjusted residuals. We used the Bonferroni method to correct for multiple testing. Statistical analyses of microbiome composition and function Alpha and beta diversity We computed alpha diversity using the Shannon entropy (diversity) of microbial abundances. We also derived species-level beta diversity using the Euclidean distance applied to species relative abundances after Centered Log-Ratio (CLR) transformation using the function vegdist of vegan R package (version 2.6_8). As CLR transformation cannot be applied to zero values, zeros in the datasets were adjusted by adding half of the lowest non-zero value to each cell. In addition, CLR-transformed abundances were adjusted for technical covariates (DNA concentration, total number of post-QC reads, and sequencing batch number) before calculating beta diversity and for all downstream analyses, using the following model: taxa_adj=residuals (taxa_CLR ~ DNA_concentration + Reads_postQC + batch) . To further characterize between-samples variation, principal coordinates analysis (PCoA) was applied to the beta diversity dissimilarity matrix within each visit using the cmdscale() function (R version 4.4.1). This approach projects samples into a reduced-dimensional Euclidean space preserving pairwise dissimilarities. To evaluate differences between phases we first calculated the average beta diversity per each woman within visits. Pairwise differences of alpha and average beta diversity were tested using Wilcoxon rank-sum tests for comparisons across centers, and using paired Wilcoxon signed-rank tests for menstrual phase comparisons. P-values were adjusted for multiple testing using Bonferroni correction. Analysis of variance To estimate the contribution of host factors (such as phases, hormone levels—E2, P4, FSH, LH, PRL, and dietary factors — nutrient 3days , FFQ, and dairy/FFQ clusters) to microbial diversity, we used permutational multivariate analysis of variance (PERMANOVA) with 9999 permutations, as implemented in the adonis2 function in the vegan R package (v. 2.6_8). First, we fitted univariate PERMANOVA models for each variable. Each model was adjusted for BMI, Bristol stool scale score, age and time of faecal sample collection, and repeated measures were accounted for by restricting permutations within individuals with the 'strata' option. To obtain effect sizes (R²) for each variable, we used marginal tests (by = 'margin'). Variables showing nominal significance (P < 0.05) in the univariate analyses were subsequently selected and included in a multivariate PERMANOVA model. This combined model incorporated all the selected variables simultaneously, adjusting for the same set of covariates (BMI, Bristol Stool Scale score, age and faecal sample collection time) and accounting for repeated sampling through subject-level stratification. Marginal effects were used to quantify the independent contribution of each variable while controlling for all the other variables in the model. Longitudinal changes of microbiome and microbiome association with diet, menstrual phases and hormones For the downstream analysis, we focused on bacterial genera, species, and pathways with a mean relative abundance greater than 0.1% and present in at least 20% of samples. This led to the inclusion of 257 taxa (71 genera and 117 species) and 185 pathways. Like bacterial abundances, microbial pathways abundances were CLR-transformed and adjusted for technical covariates. Differences in adjusted microbiome composition (microbial taxa abundances and pathways) between the three collection centers were evaluated using the Kruskal-Wallis test. To evaluate the impact of macronutrient intake on gut microbiome, we used the average intake estimated from daily diaries across 3 days prior to stool collection (nutrient 3days ). LMMs were applied for each macronutrient, adjusting for phase, age, body mass index (BMI), Bristol stool scale, recruiting center, and collection time (in the morning or during the day), with women ID included as a random effect. taxa_adj ~ phase +nutrient 3days + Age + BMI + Bristol_stool_scale + Center + Collection_when + (1 | ID) In addition to nutrient intake, associations between the gut microbiome and dietary patterns were also assessed using nutrients derived from the FFQ (see Supplementary Notes ) and diet quality scores. Furthermore, dietary patterns, identified through principal component analysis and K-means clustering of macronutrient intake proportions, were also tested. Specifically, women were grouped into clusters based on similarities in their dietary profiles (see Supplementary Notes ). Association analysis with dietary patterns was performed using LMMs with the same set of covariates described above. We investigated first associations between microbial species and menstrual phases using paired Wilcoxon signed-rank tests on CLR abundances, after adjustment for technical covariates. Then, we also carried out this analysis using LMMs to account for the longitudinal structure of the data. The model included menstrual phase (treated as a categorical variable with follicular phase as the reference), age, BMI, Bristol stool scale, recruiting center, and time of collection as fixed effects, with women ID as a random intercept. The formula used was: taxa_adj ~ phase + Age + BMI + Bristol_stool_scale + Center + Collection_when + (1 | ID) To investigate associations with reproductive hormones, we used a similar model, adding in turn each of the hormones to the model, and treating phase as numeric, continuous variable, as follows: taxa_adj ~ phase + hormone + Age + BMI + Bristol_stool_scale + Center + Collection_when + (1 | ID) For the association with dietary features, hormonal phases and sex hormones, we considered significant the results with p-values < 0.05 and Benjamini–Hochberg–corrected q < 0.25. Mediation analyses of hormones, microbiome and diet In order to investigate if hormones mediate the impact of diet on the gut microbiome, we first identified microbial species and pathways that were significantly associated with both hormone levels and diet. Only those species/pathways showing associations with both factors were considered for mediation testing, as these factors can be intermediaries through which diet and hormones influence microbial composition. Mediation analyses were performed using the mediation R package (v.4.5.1) and applying a two-model approach. First, a mediator model was fitted using linear regression, in which hormone levels were modeled as a function of dietary exposure and covariates. Secondly, an outcome model was fitted, modelling the microbiome as a function of dietary exposure and the mediator. Both models were LMMs to account for repeated measures and included the same covariates as in the analyses described above. The mediation analysis was then conducted using the mediate() function to estimate the average causal mediation effect (ACME), representing the indirect effect through the hormone mediator; the average direct effect (ADE) of diet on the microbiome, independent of the mediator, and the total effect. Statistical significance and confidence intervals were estimated using non-parametric bootstrapping with 1,000 simulations. Assessment of enzymes involved in estrobolome activity We quantified 7,976 bacterial enzyme abundances mapping the HUMAnN output to level 4 enzyme commission (EC) categories. According to HUMAnN protocol, we normalized counts by total number of postQC sequencing reads and derived relative abundances. They were CLR transformed and adjusted for technical covariates. Among those with prevalence > 20%, we selected enzymes that were potentially involved in the estrobolome , in particular enzymes involved in glycoconjugate metabolism (beta-glucuronidase, beta-galactosidase, beta-glucosidase, 6-phospho-beta-glucosidase) and sulfate metabolism (arylsulfatase type 1 and aryl-sulfate sulfotransferase) 15 , 29 , 30 , and evaluated the association between their abundances, menstrual phases, and circulating reproductive hormones levels using paired Wilcoxon signed-rank tests. We also repeated the association using LMMs; however, outliers (defined as samples with enzyme abundance > 3 standard deviation from the mean) were removed prior to model fitting. Specifically, we used the following LMM models: enzyme_adj ~ phase + Age + BMI + Bristol_stool_scale + Center+ Collection_when + (1 | ID) and enzyme_adj ~ phase + hormone + Age + BMI + Bristol_stool_scale + Center + Collection_when + (1 | ID) Results Within the Wome4Health cohort we analysed reproductive hormones, gut microbiome and dietary information across 4 visits during the menstrual cycle for 160 women not taking oral contraceptives. Participants were healthy women with a mean BMI of 21.93 (S.D. 2.9) and mean age of 29.2 years old (S.D. 2.9 years). Detailed sample characteristics are provided in Supplementary Table S1 . Consistency between dietary patterns and nutrients intake from FFQ and daily records To investigate the impact of diet on microbiome, we first evaluated consistency between two types of dietary information collected in the study: a frequency food questionnaire (FFQ) administered at enrollment, and a daily diet registry compiled by participants on the MyFitnessPal app. The latest was available only for a subset of 149 out of the 161 women. From the FFQ, we estimated the average macronutrients per day using standard portions obtained from the Italian Society of Human Nutrition’s table (SINU – Società Italiana di Nutrizione Umana, https://eng.sinu.it/ ) and nutrient data from the Italian Food Composition Database (BDA) ( https://bda.ieo.it/ ) (see Supplementary Notes ). Daily diaries underwent strict quality control to identify and correct errors in nutrients definition ( Supplementary Notes ). We observed that the average nutrient intake calculated from all daily diaries collected across the entire study period differed significantly from the estimated average macronutrient intake derived from FFQs ( Supplementary Table 2 ). Daily diaries reported, on average, significantly higher intakes of total energy, carbohydrates, fats, proteins and sugars compared with FFQs (Bonferroni-adjusted p-value < 0.001 for all comparisons). In contrast, FFQs estimated significantly higher intakes of cholesterol and fiber than daily diaries (Bonferroni-adjusted p-value 0.05). Nevertheless, measurements were highly correlated ( Supplementary Table 2 ), except for sugar and sodium, probably reflecting the difficulties in estimating sweets/snacks on FFQs. Association between diet and gut microbiome species abundance Since our data was collected along a menstrual cycle, we first evaluated the effect of the hormonal phase on each nutrient intake (using the average of the 3-day data recorded prior to fecal sample collection, averaged within the same hormonal phases, nutrients 3days , see Methods ) ( Supplementary Table 3 ). While some associations showed a nominal p-value < 0.05, none remained significant after Bonferroni correction for multiple comparisons, indicating that diet patterns are not significantly different across the menstrual cycle. We then examined how nutrients intake influenced gut microbiome species and pathways. We characterized the taxonomic and functional composition of the gut microbiome of 542 fecal samples collected from 160 women across the three centers along four menstrual phases. Shotgun metagenomic data identified, within the Bacteria kingdom, a total of 12 phyla, 303 classes, 317 orders, 353 families, 1,043 genera, 1,740 species, and 532 metabolic pathways ( Supplementary Fig. 1 ). Given our interest in fine-scale microbial variation and metabolic pathways, subsequent analyses focused primarily on lower taxonomic ranks, particularly at the species level and on microbial pathways. To reduce sparsity and increase robustness of our data, we applied stringent filtering criteria retaining only features with a relative abundance > 0.1% and present in at least 20% of samples. After filtering, 117 species and 185 pathways were retained for downstream analyses. We identified 26 associations of nutrients 3days with species abundances and 28 with microbial pathways at BH q-value < 0.25 ( Supplementary Tables 4 and 5 , Fig. 2 A, Supplementary Fig. 2 ). Among these, we see some expected associations. For example, higher fiber intake was associated with increased abundance of Roseburia hominis (effect estimate = 0.215, q = 0.112, where the estimates represent changes in standard deviation units of CLR abundance), a butyrate‑producing taxon that has been previously linked to beneficial metabolic effects and enrichment with dietary fiber consumption 31 . Likewise, a negative association between sugar intake and abundance of Clostridium sp. AM22_11AC (effect estimate = − 0.155, q = 0.060) is consistent with prior studies indicating that diets higher in simple sugars can alter gut microbiota composition, including reductions in certain Firmicutes‑associated taxa, such as Clostridium 32 . Additionally, we also found that fat intake was positively associated with Bifidobacterium bifidum (effect estimate = 0.101, q = 0.158), a species often investigated in studies related to diet. Although we observed a positive trend between fat intake and Bifidobacterium bifidum abundance, evidence from human studies indicates that the effects of dietary fat on Bifidobacterium spp. are highly context‑dependent and influenced by both the quantity and type of fat consumed 33 . Interestingly, we observed that protein intake also showed a modest negative association with Acidaminococcus intestini (effect estimate = − 0.093, q = 0.101). This bacterium is known for fermenting amino acids, such as glutamate, linking to potential roles in protein and amino acid metabolism in the gut 34 . Among microbial pathways, the strongest association was observed between higher cholesterol intake and reduced abundance of pathways involved in bacterial lipid biosynthesis, including anaerobic gondoate biosynthesis (effect estimate = − 0.045, q = 0.020) and cis-vaccenate biosynthesis (effect estimate = − 0.043, q = 0.023). Furthermore, using the same mixed-effects modeling framework applied in previous association analyses, we investigated the relationship between gut microbial species abundance and habitual food consumption based on FFQ data (weekly intake frequencies). We identified several diet–microbiome associations (54 with species and 144 with pathways), with beverage-related patterns showing the strongest signals (Fig. 2 B, Supplementary Fig. 3, Supplementary Tables 6 and 7 ). In particular, coffee consumption was strongly positively associated with a butyrate-producing Lawsonibacter asaccharolyticus (effect estimate = 0.464, q = 1.38×10⁻⁵), confirming previous large-scale population-based metagenomic studies 6 . Interestingly, Lawsonibacter asaccharolyticus also showed a positive association with yogurt consumption (effect estimate = 0.256, q = 0.063). This association was not significantly altered (effect estimate = 0.21, q = 0.10) after adjustment for coffee intake, indicating an independent relationship with yogurt consumption. This specific taxon has not been previously described among the microbiome shifts associated with yogurt intake. Additional positive associations with coffee consumption were observed for Alistipes communis and Alistipes ihumii , both members of the genus Alistipes which has been previously linked to higher caffeine consumption 35 . Another known association was observed between yogurt intake and Streptococcus thermophilus (effect estimate = 0.344, q = 0.0068), a well-established yogurt starter species 36 . Among the pathways, particularly intriguing was the association between olive oil consumption and the superpathway of beta-D-glucuronosides degradation (effect estimate = 0.106, q = 0.014), a complex microbial pathway that includes among key enzymes the beta-glucuronidase activity. This may suggest a potential link between diet, microbial deconjugation capacity, and hormone-related metabolism. Several other associations were found at a lower level of significance, but larger studies are needed to confirm their validity. Panel A shows a heatmap of associations between microbial species and nutrients intake (nutrients 3 days ). Panel B shows a heatmap of associations between microbial species and food habits based on FFQ. Only species showing at least one association at q-value < 0.25 are shown. Squares are colored according to the effect size of the association, as in the side legend. Asterisks mark level of significance (* q < 0.25, **q < 0.1, ***q < 0.05). Finally, we used several metrics of dietary patterns, specifically dietary definitions derived by clustering the observed macronutrients intake, and established diet quality indices (the Alternate Mediterranean Diet (aMED), the Healthy Eating Index (HEI) and Alternate Healthy Eating Index (aHEI), and the Dietary Approaches to Stop Hypertension (DASH score) and associated these with species abundances, alpha and beta diversity ( Supplementary Table 8, Supplementary Table 9, Supplementary Table 10 ). Alpha and beta diversity were not significantly associated with diet scores. Therefore, the overall diet habits were overall homogeneous in our data and did not drive major changes of the gut microbiome. Only a few associations were observed at the species level. Contribution of diet, hormone levels and menstrual phase to microbiome To quantify the relative influence of diet, menstrual phases and reproductive hormones on gut microbiome, we performed a series of Permutational Multivariate Analysis of Variance (PERMANOVA) analyses on species-level beta diversity (see Methods ). We modeled beta diversity as a function of menstrual phase, sex hormones, dietary patterns derived from FFQ and short-term food diary clusters. We first performed univariate PERMANOVA models to evaluate contribution of each variable while adjusting for covariates. Several dietary components and hormonal variables showed nominal associations with microbial beta diversity, although all effect sizes were small. Among dietary factors, wine (R² = 0.58%, p = 0.018), fish (R² = 0.57%, p = 0.027), olive oil (R² = 0.65%, p = 0.026), butter consumption (R² = 0.54%, p = 0.032), and fiber intake from daily diaries (R² = 0.47%, p = 0.016) reached nominal significance. Among hormonal variables, P4 (R² = 0.15%, p = 0.024), LH (R² = 0.21%, p = 0.004) and FSH (R² = 0.31%, p = 0.005) were also nominally associated with beta diversity. In contrast, menstrual phase and dietary pattern clusters derived from FFQ and short-term diaries showed no association (Supplementary Table 11) . Variables with nominal significance in univariate analyses were included in a multivariable PERMANOVA model. The full model explained 7.6% of the total variance in microbial beta diversity (p = 0.001), meaning that a large fraction of the variability remains unexplained. When estimating the independent contribution of each variable within the multivariable model using marginal tests, only fiber intake (R² = 0.46%, p = 0.017) and Bristol Stool Scale score (R² = 0.69%, p = 0.016) remained statistically significant. All other variables, such as dietary components (wine, fish, olive oil, butter) and hormones (P4, LH, FSH), were no longer significant (Supplementary Table 11) . Characterization of gut microbiome across menstrual phases Next, we focused on changes in gut microbiome during the menstrual cycle. Intriguingly, alpha diversity showed significant changes, in particular it was lower in the follicular phase compared to ovulatory phase (Bonferroni-adjusted Wilcoxon test p-value = 0.038), while other phase-to-phase comparisons were not statistically significant (all nominal p > 0.05) (Fig. 3 A, Supplementary Table 12 ). Likewise, average beta diversity per woman significantly increased from earlier phases to late luteal phase (follicular versus late luteal, Bonferroni-adjusted paired Wilcoxon test p = 8.7 × 10 − 6 ; ovulatory versus late luteal, p = 6.4 × 10 − 4 ; early luteal versus late luteal, p = 4.2 × 10 − 6 ) (Fig. 3 B, Supplementary Table 12 ). These results suggested that microbial community composition undergoes a reorganization during the menstrual cycle. We also examined the association between individual hormones and beta diversity. Spearman correlation analysis between PCoA1 of beta diversity and hormones revealed a significant positive association with FSH during the early luteal phase (Spearman correlation ρ = 0.20, p = 0.005, see Supplementary Fig. 4 ), while no significant correlations were observed in other phases or for LH, P4, PRL, and E2 . Microbiome composition was overall stable across visits. At the genus level, the highest abundant bacterial genera were Bacteroides and Phocaeicola whereas, at species level, microbiome was mostly dominated by Faecalibacterium prausnitzii, Bacteroides uniformis and Phocaeicola vulgatus ( Supplementary Fig. 5 ; Supplementary Fig. 6 ). Stratification by menstrual phases revealed only minor differences in single species abundances ( Fig. 4 ) . Slight differences at the level of species abundances were instead seen between recruiting centers ( Supplementary Fig. 7 ), and thus, the city of collection was used as covariate in all subsequent analyses. Next, we investigated associations between gut microbial species abundance and pathways with menstrual phases using first Wilcoxon signed-rank tests and then linear mixed-effects models. In the paired Wilcoxon signed-rank analysis, Bacteroides faecis was the only significant species (q‑value = 0.21), showing a difference between the follicular and early luteal phases ( Supplementary Table 13 ). In the linear mixed-effects models, in addition to B. faecis , we detected other 10 species and 5 pathways exhibiting differences in abundance between the follicular and subsequent phases ( Supplementary Table 14 ; Supplementary Table 15) . Blautia massiliensis showed increased abundance in ovulatory (effect estimate = 0.20, q = 0.18), early luteal (effect estimate = 0.28, q = 0.03) and late luteal phase (effect estimate = 0.26, q = 0.15) as compared to the follicular phase. Conversely, Lacrimispora amygdalina decreased in late luteal (effect estimate = -0.45, q = 0.04), while Roseburia intestinalis decreased in early and late luteal phases (effect estimate = -0.39, q = 0.17; and effect estimate = -0.41, q = 0.23, respectively). Other notable associations included increased abundance of Dysosmobacter welbionis (late luteal, effect estimate = 0.28, q = 0.17), Clostridium spp . (early luteal, Clostridium sp AF36 4 , effect = 0.23, q = 0.22; Clostridium sp AM22 11AC , effect estimate = 0.19, q = 0.23), and Roseburia hominis (late luteal, effect estimate = 0.42, q = 0.24) (Fig. 5 A ). Relationship between serum levels of reproductive hormones hormones and the gut microbiome Next, we evaluated the relationship between gut microbiome species and pathways abundance with the five circulating reproductive hormone levels: progesterone (P4), luteinizing hormone (LH), follicle-stimulating hormone (FSH), 17β-estradiol (E2) and prolactin (PRL), using linear mixed-effects models. A total of 28 species and 22 pathways showed significant associations with reproductive hormones ( Supplementary Table 16; Supplementary Table 17; Fig. 5 B). Interestingly, we observed the highest number of significant associations not for the most well-studied hormones such as 17β-estradiol (N = 3) and progesterone (N = 8), but rather for FSH and LH (N = 20 and 19, respectively). For example, LH and FSH were positively associated with Phocaeicola vulgatus (also known as Bacteroides vulgatus ) (q = 0.06 and q = 0.05, respectively), a bacterial species known to induce PCOS-like phenotype in mice 37 . Thus, while the association with these hormones is intriguing, the effect directions are not always consistent with the physiology of PCOS patients, characterized by high LH and low-to-normal FSH. Furthermore, these hormones were also positively associated with another species of the same genus Phocaeicola massiliensis (q = 0.08 and q = 0.08, respectively), for which no evidence exists, to our knowledge, for an involvement in gynecological or endocrine disorders, and with Faecalibacterium prausnitzii (q = 0.23 and q = 0.06, respectively), for which association with PCOS is ambiguous 38 , 39 . Associations with 17β-estradiol highlighted a negative link with Akkermansia muciniphila (q = 0.22), albeit this is in contrast with studies carried out in vitro showing an increase of this species after estradiol supplementation 40 . Notably, the effect size of all the significant associations between menstrual phases, reproductive hormones, bacterial species and pathways, remained consistent even when adjusting the model with food habits from FFQs or nutrients 3days from daily diet ( Supplementary Fig. 8 ). Diet directly shapes gut microbiome taxa and pathway abundance independent of hormonal mediation To investigate whether hormonal fluctuations mediate the impact of diet on the gut microbiome, we considered species that were significantly associated with both circulating hormone levels and short-term dietary intake (nutrients 3days ), as these taxa represent plausible intermediates through which diet and hormones may influence microbial composition. Five bacterial species were included in the mediation analysis: Alistipes shahii (linked to prolactin and dietary cholesterol), Bifidobacterium bifidum (linked to LH, FSH, and dietary fat), GGB9758_SGB15368 (linked to FSH and dietary fat), Mediterraneibacter faecis (linked to LH and dietary sugar), and Phascolarctobacterium faecium (linked to FSH and dietary sugar). We also applied the same approach to microbial functional pathways to explore any potential mediation by hormones. The pathway PWY0.1479 (tRNA processing) was associated with FSH and LH, as well as with dietary intake of calories, fiber, and proteins, and was therefore included in the mediation analysis. The average causal mediation effect (ACME) was not statistically significant (all p-values > 0.33), indicating that in our data hormones did not mediate the association between diet and bacterial abundance. The average direct effect (ADE) of diet on bacterial abundance remained significant for most bacteria (e.g., B. bifidum , ADE p-value = 0.004–0.01; M. faecis , ADE p-value = 0.018; P. faecium , ADE p-value = 0.002), and the total effect was also significant. The proportion of the total effect mediated through hormones was non-significant ( Supplementary Table 18 ). Likewise, for the tested pathway (PWY0-1479: tRNA processing), the ACME was not statistically significant, with confidence intervals consistently crossing zero. In contrast, the ADE of diet on pathway abundance remained significant for multiple exposures, such as protein, calorie, and fiber intake (ADE p-value = 0.004–0.016), and the total effect of diet on pathway was also significant. The proportion of the total effect mediated via hormones was also modest (ranging from ~ 0% to − 0.3%) and non-significant ( Supplementary Table 19 ). These findings indicated that the associations between short-term dietary intake and both gut bacterial abundances and metabolic pathways are predominantly direct, rather than mediated by hormonal fluctuations. This supports the concept in which diet directly shapes microbial composition and functional capacity, independent of circulating hormone levels. Estrobolome enzyme abundance across menstrual phases Many of the bacteria we found to be associated with hormones, possess genes encoding the “estrobolome” - the collection of microbial enzymes hypothesized to participate in human estrogen metabolism, with a potential for progesterone metabolism. Therefore, we thought to specifically evaluate the abundance of six microbiome enzymes known or hypothesized to be involved in the estrobolome activity, including established enzymes such as beta-glucuronidase and beta-galactosidase, and several potentially relevant enzymes: arylsulfatase, aryl-sulfate sulfotransferase, 6-phospho-beta-glucosidase and beta-glucosidase ( Supplementary Table 20) . We quantified the total microbiome abundance of each enzyme across the four menstrual phases (see Methods ) using HUMAnN (Supplementary Fig. 9) . We found that throughout all phases, beta-galactosidase was the most prevalent, while aryl sulfate sulfotransferase exhibited the lowest abundance ( Supplementary Fig. 9 ). We then examined the gene abundance differences across pairs of menstrual phases using paired Wilcoxon signed-rank tests ( Supplementary Fig. 10) , and observed that after multiple testing correction there were significantly lower levels of beta-galactosidase, beta-glucosidase, arylsulfatase and beta-glucuronidase in follicular versus early luteal phase (q-values 0.01, 0.05, 0.13 and 0.22, respectively). Furthermore, beta-galactosidase abundance was also lower in the ovulatory phase compared to follicular phase (q = 0.22) (Supplementary Table 21). Next, we evaluated the association between enzyme abundances and menstrual phases, and between enzyme abundances and reproductive hormones (17β-estradiol and progesterone) using linear mixed-effects models. However, none of the enzymes showed significant change across phases ( Supplementary Table 22 ) or associations with hormones ( Supplementary Table 23 ). To evaluate the overall contribution of enzyme abundance changes to reproductive hormone levels, we compared the performance of a basic model that included progesterone (or 17β-estradiol ) as response variable and host covariates as explanatory variables (menstrual phase, BMI, age, Bristol stool scale, time of sample collection and technical covariates) with a complete model that incorporated these variables as well as the six enzyme abundances. For progesterone, the complete model did not improve the fit compared to the base model (p = 0.56). Moreover, model selection criteria indicated no improvement, with higher AIC (1492 vs 1476) and BIC (1562 vs 1536) values observed for the complete model. Similarly, for 17β-estradiol, the inclusion of the gene enzymes did not improve the model performance (p = 0.55). Again, the complete model showed higher AIC (1116 vs 1103) and BIC (1199 vs 1154) values compared to the base model. Discussion There is currently a significant knowledge gap regarding gut microbiome composition and dynamics across a natural menstrual cycle in women of reproductive age without gynecological or metabolic conditions. The present study addressed this gap while additionally examining dietary parameters and reproductive hormone fluctuations during the menstrual cycle. To assess the contribution of diet to microbiome variability, we first evaluated the consistency in dietary intake estimates obtained by two sources of dietary information: food frequency questionnaires (FFQs) and daily food diaries. Comparison of nutrients estimated by these two approaches revealed significant discrepancies consistent with known bias in subjective surveys 41 . In particular, daily food diaries - which likely provide a more accurate reflection of actual food consumption compared to a retrospective assessment - provided higher estimates of total energy and macronutrients, but lower estimates of cholesterol and fiber compared to the FFQ. These differences may arise from several aspects, including limited dietary self-awareness, the simplified FFQ used in this study, or potential errors and lack of country-specific data in the MyFitnessPal database as previously reported 42 , 43 . Nevertheless, despite absolute differences in nutrient estimates, nutrients derived from the two methods showed significant positive correlation. This indicates a generally consistent ranking of individuals across assessment approaches, supporting the complementary use of FFQs to capture habitual food consumption and daily food diaries to reflect nutrient intake proximal to microbiome sampling. We next examined associations between dietary components and gut microbiome composition and function, using dietary parameters obtained from both types of information. Several of the observed associations replicate previously reported findings, reinforcing the validity of our analyses. For example, we confirmed the positive association between fiber and fruit intake with abundance of Roseburia hominis , a known short-chain fatty acid (SCFA) producer, capable of breaking down dietary fibers 44 – 46 . We also confirmed the well-established positive association between coffee consumption and Lawsonibacter asaccharolyticus 6 . At the same time, though, we observed a novel association of L . asaccharolyticus with yogurt intake that was independent of coffee consumption, suggesting that this species may respond more broadly to fermented food consumption. Furthermore, we identified additional species belonging to the Alistipes genus (e.g. A. ihumii and A. communis ) to be associated with coffee consumption, a genus that has previously been associated with coffee and caffeine intake 35 . Among the novel findings, Alistipes shahii and Hominenteromicrobium mulieris were associated with dietary cholesterol intake. Considering that the genus Alistipes is receiving growing attention in dyslipidemic conditions, and that H. mulieris has been recently associated with intestinal mucosal health 47 , these associations merit further investigation to clarify whether high dietary cholesterol selectively promotes microbial taxa with potential beneficial or compensatory functions. Another novel relationship involved Parabacteroides merdae , a species known to metabolize branched-chain amino acids (BCAAs) 48 , which was associated with cold cuts consumption, foods typically rich in BCAA. Diet also influenced microbiome function, which we capture using microbial pathways abundances derived from our shot-gut sequencing approach. Among worth noting associations, is the inverse relationship between cholesterol intake and pathways involved in bacterial lipid biosynthesis. Consistent with this observation, prior animal studies have reported correlations between fasting triglyceride levels and bacterial lipid-related pathways, including CMP 3 deoxy D manno-octulosonate biosynthesis 49 . Collectively, these findings reinforce the notion that both individual foods and general dietary patterns influence gut bacterial composition and function. Sex is another host factor consistently associated with gut microbiome interindividual variability. Several studies have shown that young women show greater microbial diversity than age-matched men; however this difference diminishes with age and is no longer detectable after menopause 10 , 11 , 16 , 50 . This pattern has been largely attributed to the decline in estrogen production during menopause and its potential interaction with gut bacteria involved in estrogen metabolization, collectively referred to as the estrobolome 16 . Microbial taxa shown to decrease after menopause include species from Prevotella, Bacteroides, Parabacteroides ., as well as A. muciniphila , and their abundance strongly correlated with depleted gut microbial β-glucuronidases enzymes 16 . Despite this evidence, the contribution of other reproductive hormones, many of which undergo substantial changes during menopause, remain largely unexplored in both pre- and post-menopause women. In this context, we assessed the impact of menstrual cycle phases and five reproductive hormones on the gut microbiome. Intriguingly, microbiome changes were visible and significant already on overall microbiome metrics such as nd richness and diversity. Richness was modestly higher during the ovulatory phase compared to the follicular phase. Differences in beta diversity were instead more pronounced, with the greatest interindividual variation observed during the late luteal phase compared to the follicular phase. This suggests that proximity to menstruation may be associated with increased microbial variability between individuals. At higher resolution, we identified microbial species and microbial pathways whose abundances were associated with a menstrual phase or with circulating reproductive hormones or both. Some of these associations align with previous findings on patients, thus bringing additional support to the detected relationships. For example, progesterone levels were positively associated with Collinsella aerofaciens . Abundance of this species was previously shown to increase in endometriosis patients following treatment with Dienogest, a progestin medication 51 . Intriguingly, most of our associations were detected with FSH, LH and progesterone rather than with 17β-estradiol. The association with FSH and LH is likely indirect, potentially via the gut-brain axis, signaling to hypothalamus and pituitary glands 52 . The gut-brain axis may also explain the few associations between microbial species and pathways with prolactin levels, prolactin being secreted under inhibitory control by hypothalamic dopaminergic neurons. Finally, a subset of species associated with hormones were also associated with dietary factors, but we were unable to find evidence for mediation of either diet on gut-hormones associations or of hormones on diet-gut associations. Despite only a few associations being observed with 17β-estradiol, we noted that several included bacterial species potentially involved in the estrobolome activity, particularly members of the Bacteroides genus. While the underlying mechanisms of estrobolome have been widely described for estrogens, they could potentially be extended to progesterone 53 . We thus quantified the abundance of genes encoding estrobolome enzymes and investigated their association with circulating 17β-estradiol and progesterone levels. Intriguingly, we see menstrual cycle associated changes in abundance of the most relevant enzymes: beta-galactosidase, beta-glucosidase, and higher arylsulfatase type1 and beta-glucuronidase. None of these were however associated with circulating hormones, a negative result that may be attributed to the fact that we are unable to distinguish systemic and de-conjugated fractions, which may more directly reflect microbial activity, or to the fact that we measured hormones in plasma rather than in feces. Indeed, while mechanistic studies suggest the steroid metabolizing capacity of the gut microbiome, human studies have so far yielded inconsistent results regarding association between estrobolome enzymes and plasma circulating estrogen levels 54 , 55 . In parallel, our data highlight the importance of SCFA-producing bacteria in the context of hormonal regulation. In fact, a total of 21 out of the 30 species associated with hormone and diet are also SCFA producers, including the well-characterized Akkermansia muciniphila and Faecalibacterium prausnitzii . While beta-glucuronidase producing bacteria facilitate estrogen recirculation via enterohepatic recycling, we hypothesize that SCFA producing taxa may exert complementary effects by first modulating host metabolism and, in turn, impact endocrine responses and hormonal homeostasis. Direct measurement of SCFA also in future expansions of this study are needed to clarify this hypothesis. To our knowledge, this study provides the first evidence of microbiome variations during menstrual cycle in healthy women with concurrent hormone measurements, and the first to interrogate these changes in relation to reproductive hormones beyond estrogen and progesterone. We also acknowledge its limitations. First, the sample size is still relatively small, and thus we may be underpowered to pinpoint smaller effects of hormones on microbiome. Furthermore, our study includes only healthy women, in whom physiological hormonal changes occur, so microbiome–hormone interactions are likely too subtle to fully be captured. Second, we measured only total circulating hormones, without performing specific assays to differentiate between conjugated forms, directly synthesized unconjugated forms, or unconjugated forms resulting from deconjugation. Third, volunteers exhibited relatively homogeneous dietary patterns, limiting our power to detect potential hormonal mediation of diet–microbiome interactions. Finally, we acknowledge that other unmeasured factors, such as sleep quality and stress, could also influence microbiome dynamics during the menstrual cycle. In conclusion, our findings indicate that physiological fluctuation in endogenous reproductive hormones across a natural menstrual cycle are associated with measurable changes in gut microbiome composition and function, and their effects are distinct from dietary factors. These results underscore the importance of incorporating sex-specific data and hormonal status into microbiome research. Our study provides novel insights into host-microbiome interactions that are specific to female physiology, highlighting a strong relationship with gonadotropics hormones. Future research will be necessary to build upon these observations and clarify their clinical relevance, especially for reproductive health. In particular, larger longitudinal cohorts integrating multi-omics approaches will be critical to elucidate how gonadotropins together with sex steroids interact with microbial metabolism and host signaling pathways. Such efforts will help determine whether hormonally-associated microbiome changes translate into measurable physiological consequences and whether they can be leveraged to inform personalized, menstrual cycle-aware interventions aimed at improving women’s health. Supplementary Notes File (PDF) In the Supplementary Notes we describe how dietary data were collected, processed, and analyzed in the study. The file provides details on the transformation of FFQ data into nutrient intakes and diet scores, and reports differences in nutrient intake, diet scores, and gut microbiome composition across centers and hormonal phases. Supplementary Notes figures and tables are included along with their legend. Supplementary Figures (PDF) In the Supplementary Figures file we provide Supplementary Fig. 1 to 10. Supplementary Figures legends are included in the file. Supplementary Tables (xlsx) This file contains 23 Supplementary Tables and one tab containing a summary. Data availability statement All our code used to analyse the data and instructions are available at: https://github.com/Sanna-s-LAB/Women4Health/tree/main/Code_for_LoFaro_et_al We provide relative abundance of species and pathways in Zenodo along with information on visit number and menstrual phase at the following DOI: 10.5281/zenodo.19953398 . The data will be made public after article acceptance. The temporary link for reviewers is the following: https://zenodo.org/records/19953398?preview=1&token=eyJhbGciOiJIUzUxMiJ9.eyJpZCI6IjRmYjljODEzLTRiZTItNDVhMi05MjY4LWY1M2UyMmVlNGY1NSIsImRhdGEiOnt9LCJyYW5kb20iOiIwZTQwN2QzZTNlZjYyMzNkMTMzNDg4NmRkZmYyZGJhZCJ9.8wKj4f4cK-eZS1DEnbaiLOqFL-wt4L3LTJjpTz5j3vBvCT6t4Jt4iGfGIq6POgtmzBz8txRrnVyvCv5FzgsHpg . Other participant-level personally identifiable data cannot be shared in respect to participants' informed consent and are protected under the Italian Personal Data Protection Code and European Regulation 2016/679 of the European Parliament and of the Council (GDPR) that prohibit distribution even in pseudo-anonymised form. All data necessary to support the conclusions drawn in this study are available in the Supplementary Tables. Declarations The study protocol - including informed consent forms - was approved by the Friuli-Venezia Giulia Ethical Committee on 09/2021 with prot. N. 0034184/P/GEN/ARCS and modifications amended on 09/2023 with prot. N. 0033477/P/GEN/ARCS; by the Sardinian Ethical Committee on 05/23 with prot. N. 1.1 and modifications amended on 07/2024 with prot. N.1.2; and by the Area Vasta Emilia Romagna Ethical Committee on 12/2024 with prot. N 1875/2024. Funding This study was co-funded by the European Union (ERC Stg 2022 to S.S., acronym SEMICYCLE, GA n.101075624), by the Next Generation EU, in the context of the National Recovery and Resilience Plan, Investment PE8 – Project Age-It: “Ageing Well in an Ageing Society ” [DM 1557 11.10.2022 to S.S.], by NutrAGE grant (CNR Project FOE-2021 DBA.AD005.225 to S.S.), by InvAT grant (CNR Project FOE-2022 DSB.AD006.371.001 to S.S.) and by the Italian Ministry of Health through the contribution given to the Institute for Maternal and Child Health IRCCS Burlo Garofolo, Trieste - Italy (SD 02/21 to G.G.). Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Council. Neither the European Union nor the granting authority can be held responsible for them. In addition, S.S. also received a PRIN2022 grant from the Next Generation EU funds (DSB.PN004.021 2022PMZKEC_LS2_PRIN2022 SANNA, CUP B53D23008300006). Use of AI statement The authors declare that they have used generative artificial intelligence, ChatGPT, Quillbot (2026) and Copilot online platforms only to check spelling, grammar and expressions, adapting it to an academic scientific style. Funding acquisition: G.G., S.S. All authors read and revised the manuscript and approved its final version. Furthermore, all authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Authors’ contributions Statistical and Bioinformatic analyses: V.LoF., P.F., A.C., D.V.Z., S.S. Acknowledgements We warmly thank all the volunteers who participated in the study for their time and commitment. We also thank all collaborators involved in the recruitment phase, in particular: Dr. Giovanni di Lorenzo and the other physicians of the S.C. Ostetricia e Ginecologia of the IRCCS Burlo-Garofolo, Dr. Daniela Mazzà of the S.C. Genetics of the IRCCS Burlo-Garofolo, nurses Maria Paola Fercia and Maria Paola Orani, technician Siro Agus and Dr. Elena Poma at the Azienda Ospedaliero Universitaria di Cagliari, and Alice Cordani, Francesca Bianco and Maria Chiara Matteucci at the University of Bologna. More, we thank the Directors of IRCCS Burlo-Garofolo and of IRGB-CNR for logistic support, the Agenzia Regionale Sardegna Ricerche for providing access to laboratories, Dr Stefania Olla from IRGB-CNR for her critical feedback on the enzymes analysis, technologists Marco Masala and Michele Marongiu from the IRGB-CNR for IT support. Finally, we gratefully acknowledge our colleague, Maria Laura Ferrando, who passed away during the preparation of this work and contributed to it with enthusiasm until the end. References Carlino N et al (2024) Unexplored microbial diversity from 2,500 food metagenomes and links with the human microbiome. Cell 187:5775–5795e15 Asnicar F et al (2021) Microbiome connections with host metabolism and habitual diet from 1,098 deeply phenotyped individuals. Nat Med 27:321–332 Asnicar F et al (2026) Gut micro-organisms associated with health, nutrition and dietary interventions. Nature 650:450–458 Campaniello D et al (2022) How Diet and Physical Activity Modulate Gut Microbiota: Evidence, and Perspectives. Nutrients 14:2456 Fackelmann G et al (2025) Gut microbiome signatures of vegan, vegetarian and omnivore diets and associated health outcomes across 21,561 individuals. Nat Microbiol 10:41–52 Manghi P et al (2024) Coffee consumption is associated with intestinal Lawsonibacter asaccharolyticus abundance and prevalence across multiple cohorts. Nat Microbiol 9:3120–3134 Falony G et al (2016) Population-level analysis of gut microbiome variation. Science 352:560–564 Gacesa R et al (2022) Environmental factors shaping the gut microbiome in a Dutch population. Nature 604:732–739 Ferrando ML, Busonero F, Crobu F, Sanna S (2026) Aging in women – The microbiome perspective. Ageing Res Rev 113:102950 Zhang X et al (2021) Sex- and age-related trajectories of the adult human gut microbiota shared across populations of different ethnicities. Nat Aging 1:87–100 de la Cuesta-Zuluaga J et al (2019) Age- and Sex-Dependent Patterns of Gut Microbial Diversity in Human Adults. mSystems 4:e00261–e00219 Dabek M, McCrae SI, Stevens VJ, Duncan SH, Louis P (2008) Distribution of beta-glucosidase and beta-glucuronidase activity and of beta-glucuronidase gene gus in human colonic bacteria. FEMS Microbiol Ecol 66:487–495 Ervin SM et al (2020) Structural Insights into Endobiotic Reactivation by Human Gut Microbiome-Encoded Sulfatases. Biochemistry 59:3939–3950 Ervin SM et al (2019) Gut microbial β-glucuronidases reactivate estrogens as components of the estrobolome that reactivate estrogens. J Biol Chem 294:18586–18599 Larnder AH, Manges AR, Murphy RA (2025) The estrobolome: Estrogen-metabolizing pathways of the gut microbiome and their relation to breast cancer. Int J Cancer 157:599–613 Peters BA et al (2022) Menopause Is Associated with an Altered Gut Microbiome and Estrobolome, with Implications for Adverse Cardiometabolic Risk in the Hispanic Community Health Study/Study of Latinos. mSystems 7, e0027322 Ley D, Saha S (2025) Menopause and gastrointestinal health and disease. Nat Rev Gastroenterol Hepatol 22:556–569 Cross T-WL et al (2024) Gut microbiome responds to alteration in female sex hormone status and exacerbates metabolic dysfunction. Gut Microbes 16:2295429 Bermingham KM et al (2022) Menopause is associated with postprandial metabolism, metabolic health and lifestyle: The ZOE PREDICT study. eBioMedicine 85 Park MG, Cho S, Oh MM (2023) Menopausal Changes in the Microbiome—A Review Focused on the Genitourinary Microbiome. Diagnostics 13 Brito J et al (2025) Hormonal birth control is associated with altered gut microbiota β-diversity in physically active females across the menstrual cycle: a pilot trial. J Appl Physiol 138:739–745 Li VW et al (2024) Mass spectrometric profiling of estrogen and estrogen metabolites in human stool and plasma partially elucidates the role of the gut microbiome in estrogen recycling. 08.07.606543 Preprint at https://doi.org/10.1101/2024.08.07.606543 (2024) Vandeputte D et al (2021) Temporal variability in quantitative human gut microbiome profiles and implications for clinical research. Nat Commun 12:6740 Busonero F et al (2024) The Women4Health cohort: a unique cohort to study women-specific mechanisms of cardio-metabolic regulation. Eur Heart J Open 4:oeae012 Sanna S et al (2026) Longitudinal plasma proteomics along a menstrual cycle highlights the regulatory effect of sex hormones on inflammatory and cardiometabolic circuits. Preprint at. https://doi.org/10.21203/rs.3.rs-9148671/v1 Langmead B, Salzberg SL (2012) Fast gapped-read alignment with Bowtie 2. Nat Methods 9:357–359 Blanco-Míguez A et al (2023) Extending and improving metagenomic taxonomic profiling with uncharacterized species using MetaPhlAn 4. Nat Biotechnol 41:1633–1644 Beghini F et al (2021) Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. eLife 10, e65088 Hu S et al Gut microbial beta-glucuronidase: a vital regulator in female estrogen metabolism. Gut Microbes 15, 2236749 Godse R, Fernandes JM, Kulkarni R (2025) Characterization of β-glucosidase activity of a Lactiplantibacillus plantarum 6-phospho-β-glucosidase. Appl Microbiol Biotechnol 109:86 Huang W et al (2025) Roseburia hominis improves host metabolism in diet-induced obesity. Gut Microbes 17:2467193 Zhang Y, Walker RW, Kaplan RC, Qi Q (2025) Added sugars, gut microbiota, and host health. Gut Microbes 17:2592431 Yang Q et al (2020) Role of Dietary Nutrients in the Modulation of Gut Microbiota: A Narrative Review. Nutrients 12:381 Jumas-Bilak E et al (2007) Acidaminococcus intestini sp. nov., isolated from human clinical samples. Int J Syst Evol Microbiol 57:2314–2319 Dai A et al (2023) The Association between Caffeine Intake and the Colonic Mucosa-Associated Gut Microbiota in Humans-A Preliminary Investigation. Nutrients 15:1747 Le Roy CI et al (2022) Yoghurt consumption is associated with changes in the composition of the human gut microbiome and metabolome. BMC Microbiol 22:39 Yun C et al (2024) The microbial metabolite agmatine acts as an FXR agonist to promote polycystic ovary syndrome in female mice. Nat Metab 6:947–962 Torres PJ et al (2018) Gut Microbial Diversity in Women With Polycystic Ovary Syndrome Correlates With Hyperandrogenism. J Clin Endocrinol Metab 103:1502–1511 Chu W et al (2020) Metagenomic analysis identified microbiome alterations and pathological association between intestinal microbiota and polycystic ovary syndrome. Fertil Steril 113:1286–1298e4 Sakamuri A et al (2023) Sex hormones, sex chromosomes, and microbiota: Identification of Akkermansia muciniphila as an estrogen-responsive microbiota. Microbiota Host 1:e230010 Shim J-S, Oh K, Kim HC (2014) Dietary assessment methods in epidemiologic studies. Epidemiol Health 36:e2014009 Teixeira V, Voci SM, Mendes-Netto RS (2018) & da Silva, D. G. The relative validity of a food record using the smartphone application MyFitnessPal. Nutr Diet 75:219–225 Tosi M et al (2021) Accuracy of applications to monitor food intake: Evaluation by comparison with 3-d food diary. Nutrition 84:111018 Song L et al (2022) Roseburia hominis Alleviates Neuroinflammation via Short-Chain Fatty Acids through Histone Deacetylase Inhibition. Mol Nutr Food Res 66:e2200164 Tamanai-Shacoori Z et al (2017) Roseburia spp.: a marker of health? Future Microbiol 12:157–170 Segev T et al (2026) Diet-microbiome associations in 10,068 individuals from the Human Phenotype Project to guide personalized nutrition. Nat Med. https://doi.org/10.1038/s41591-026-04312-x Cuív PÓ et al (2025) A human microbiome-derived therapeutic for ulcerative colitis promotes mucosal healing and immune homeostasis. 04.18.25325957 Preprint at https://doi.org/10.1101/2025.04.18.25325957 (2025) Wang H et al (2025) Gut microbiota metabolism of branched-chain amino acids and their metabolites can improve the physiological function of aging mice. Aging Cell 24:e14434 Horne RG et al (2020) High Fat-High Fructose Diet-Induced Changes in the Gut Microbiota Associated with Dyslipidemia in Syrian Hamsters. Nutrients 12:3557 Santos-Marcos JA et al (2018) Influence of gender and menopausal status on gut microbiota. Maturitas 116:43–53 Pronina V et al (2025) Gut Microbial Composition on Dienogest Therapy in Patients with Endometriosis. Microbiol Res 16 Organski AC, Jorgensen JS, Cross T-WL (2021) Involving the life inside: The complex interplay between reproductive axis hormones and gut microbiota. Curr Opin Endocr Metabolic Res 20:100284 Chapadgaonkar SS, Bajpai SS, Godbole MS (2023) Gut microbiome influences incidence and outcomes of breast cancer by regulating levels and activity of steroid hormones in women. Cancer Rep (Hoboken) 6:e1847 Larnder AH et al (2026) Gut microbiome and metabolome reveal hormone-related and functional alterations in ER-positive breast cancer: a case–control study. 02.06.26345778 Preprint at https://doi.org/10.64898/2026.02.06.26345778 (2026) Nannini G, Cei F, Amedei A (2025) Unraveling the Contribution of Estrobolome Alterations to Endometriosis Pathogenesis. Curr Issues Mol Biol 47 Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryNotesv2.pdf Supplementary Notes SupplementaryTablesv4.xlsx Supplementary Tables SupplementaryFiguresv3.pdf Supplementary figures Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9888793","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":653607297,"identity":"9a194ff9-c858-4318-af0d-45a2d0d718e1","order_by":0,"name":"Serena Sanna","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA40lEQVRIiWNgGAWjYJCCA4wNEkCK+QCQkJAhRQtbAkgLD3HWMDaASB4DMElQNX/76cQDP3dY2PXP7vn86kaNBQ8D++GjG/BpkTiTu+Fg7xmJ5Bl3zm6zzjkGdBhPWtoN/B7J3XCYsU0imeFG7jbjHDagFgkeM7xa5M+/hWiRv5HzzDjnHxFaDG5AbLEzuJHD/Di3jQgthjfeAv3SJpFgeCPNjDm3T4KHjZBf5M7nbv7ws63OXu5G8uPPOd/q5PjZDx/D730oSGwAxiUoQhnYiFEOAvZAzPyBWNWjYBSMglEwsgAA7wBN7ZbOgcYAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0002-3768-1749","institution":"Institute for Genetic and Biomedical research (IRGB), National Research Council (CNR)","correspondingAuthor":true,"submittingAuthor":false,"prefix":"","firstName":"Serena","middleName":"","lastName":"Sanna","suffix":""},{"id":653607298,"identity":"37bc7352-aa5e-45b0-a92c-c937a1208622","order_by":1,"name":"Valeria Lo Faro","email":"","orcid":"","institution":"Institute for Genetic and Biomedical research (IRGB), National Research Council (CNR)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Valeria","middleName":"Lo","lastName":"Faro","suffix":""},{"id":653607299,"identity":"e27346fc-b933-4025-95a0-ef23cd6fd6b1","order_by":2,"name":"Paola Forabosco","email":"","orcid":"","institution":"Institute for Genetic and Biomedical research (IRGB), National Research Council (CNR)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Paola","middleName":"","lastName":"Forabosco","suffix":""},{"id":653607300,"identity":"8f21ec06-2965-496c-8a2b-c749f31ac3e1","order_by":3,"name":"Francesca Crobu","email":"","orcid":"","institution":"Institute for Genetic and Biomedical research (IRGB), National Research Council (CNR)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Francesca","middleName":"","lastName":"Crobu","suffix":""},{"id":653607301,"identity":"789f6019-b5ac-4a90-8464-afd9afc37423","order_by":4,"name":"Andrea Carta","email":"","orcid":"","institution":"Institute for Genetic and Biomedical research (IRGB), National Research Council (CNR)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Andrea","middleName":"","lastName":"Carta","suffix":""},{"id":653607302,"identity":"0330d15d-5040-4a66-ade2-d0d8e64547cb","order_by":5,"name":"Fabio Busonero","email":"","orcid":"","institution":"Institute for Genetic and Biomedical research (IRGB), National Research Council (CNR)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Fabio","middleName":"","lastName":"Busonero","suffix":""},{"id":653607303,"identity":"0eca14a0-95f9-430c-8d56-f91218a22893","order_by":6,"name":"Andrea Maschio","email":"","orcid":"","institution":"Institute for Genetic and Biomedical research (IRGB), National Research Council (CNR)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Andrea","middleName":"","lastName":"Maschio","suffix":""},{"id":653607304,"identity":"3f2ab500-a90e-4118-be45-5e985c99e0c4","order_by":7,"name":"Myriam Viscito","email":"","orcid":"","institution":"University of Bologna","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Myriam","middleName":"","lastName":"Viscito","suffix":""},{"id":653607305,"identity":"290150b5-91c4-44fe-ab31-01730a5fdb91","order_by":8,"name":"Stefania Lenarduzzi","email":"","orcid":"https://orcid.org/0000-0001-8450-1694","institution":"Institute for Maternal and Child Health - IRCCS","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Stefania","middleName":"","lastName":"Lenarduzzi","suffix":""},{"id":653607306,"identity":"ee09ac2b-91a4-4627-80f1-8269490cc2a9","order_by":9,"name":"Giulia Beltrame Vriz","email":"","orcid":"","institution":"Department of Medicine, Surgery and Health Sciences, University of Trieste","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Giulia","middleName":"Beltrame","lastName":"Vriz","suffix":""},{"id":653607307,"identity":"11fd57ea-6d5d-46fd-b93f-3a2c6a3c0061","order_by":10,"name":"Aurora Badalini","email":"","orcid":"","institution":"Institute for Maternal and Child Health—IRCCS \"Burlo Garofolo\"","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Aurora","middleName":"","lastName":"Badalini","suffix":""},{"id":653607308,"identity":"b141a81c-8299-48e5-95a5-d5cc3bb9ae5e","order_by":11,"name":"Maria Ferrando","email":"","orcid":"","institution":"Institute for Genetic and Biomedical research (IRGB), National Research Council (CNR)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Maria","middleName":"","lastName":"Ferrando","suffix":""},{"id":653607309,"identity":"dc01ee51-d13f-43bc-8564-f180919341d3","order_by":12,"name":"Fabio Chillotti","email":"","orcid":"https://orcid.org/0009-0005-4788-5912","institution":"Institute for Genetic and Biomedical research (IRGB), National Research Council (CNR)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Fabio","middleName":"","lastName":"Chillotti","suffix":""},{"id":653607310,"identity":"099cece4-5682-453d-a2dd-78a8da3388fe","order_by":13,"name":"Simona Incollu","email":"","orcid":"","institution":"Institute for Genetic and Biomedical research (IRGB), National Research Council (CNR)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Simona","middleName":"","lastName":"Incollu","suffix":""},{"id":653607311,"identity":"fa978f37-6f36-4ee2-b537-d61e1f29f1ef","order_by":14,"name":"Davide Murrau","email":"","orcid":"","institution":"Institute for Genetic and Biomedical research (IRGB), National Research Council (CNR)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Davide","middleName":"","lastName":"Murrau","suffix":""},{"id":653607312,"identity":"444fe464-d896-40f1-9e20-3783b3ea6997","order_by":15,"name":"Francesco De Seta","email":"","orcid":"","institution":"Institute for Maternal and Child Health - IRCCS \"Burlo Garofolo\"","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Francesco","middleName":"","lastName":"De Seta","suffix":""},{"id":653607313,"identity":"7ba856bf-cc20-492c-a8a2-16482cd47f2f","order_by":16,"name":"Giuseppe Ricci","email":"","orcid":"","institution":"Obstetrics and Gynecology, Institute for maternal and child health “IRCCS Burlo Garofolo”, Via dell’Istria 65, 34137 Trieste, Italy","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Giuseppe","middleName":"","lastName":"Ricci","suffix":""},{"id":653607314,"identity":"12b7565b-c341-4fc0-9a9a-3340eb07b4ad","order_by":17,"name":"Giovanni Barbara","email":"","orcid":"","institution":"University of Bologna","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Giovanni","middleName":"","lastName":"Barbara","suffix":""},{"id":653607315,"identity":"6ff35715-dc22-4ce1-b0db-613aaffcc147","order_by":18,"name":"Stefano Guerriero","email":"","orcid":"","institution":"University of Cagliari","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Stefano","middleName":"","lastName":"Guerriero","suffix":""},{"id":653607316,"identity":"0a3213a0-b5ef-4a25-9a3a-45c58adfd489","order_by":19,"name":"Alessandra Meloni","email":"","orcid":"","institution":"Azienda Ospedaliero Universitaria di Cagliari","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Alessandra","middleName":"","lastName":"Meloni","suffix":""},{"id":653607317,"identity":"92399014-feec-43a7-94e3-b6c550c0864e","order_by":20,"name":"Maria Barbaro","email":"","orcid":"","institution":"University of Bologna","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Maria","middleName":"","lastName":"Barbaro","suffix":""},{"id":653607318,"identity":"67b66583-9cf6-4f2f-a768-46ecfed8c477","order_by":21,"name":"Giorgia Girotto","email":"","orcid":"https://orcid.org/0000-0003-4507-6589","institution":"Institute for Maternal and Child Health - IRCCS \"Burlo Garofolo\"","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Giorgia","middleName":"","lastName":"Girotto","suffix":""},{"id":653607319,"identity":"b42c3c42-603d-4073-9eff-556150e86640","order_by":22,"name":"Daria Zhernakova","email":"","orcid":"","institution":"Institute for Genetic and Biomedical research (IRGB), National Research Council (CNR)","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Daria","middleName":"","lastName":"Zhernakova","suffix":""},{"id":653607320,"identity":"414d8fe1-a669-4e77-9edb-39d31739e9bc","order_by":23,"name":"Silvia Camarda","email":"","orcid":"","institution":"Department of Medicine, Surgery and Health Sciences, University of Trieste","correspondingAuthor":false,"submittingAuthor":false,"prefix":"","firstName":"Silvia","middleName":"","lastName":"Camarda","suffix":""}],"badges":[],"createdAt":"2026-06-01 18:04:41","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9888793/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9888793/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":112251898,"identity":"6ca761b4-d74c-40ec-824f-dc1f8cd55ac7","added_by":"auto","created_at":"2026-06-16 15:20:57","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":368500,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eSample collection scheme\u003c/strong\u003e\u003c/p\u003e","description":"","filename":"11.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9888793/v1/ea1ed593c1e2bb62d4a32eff.jpg"},{"id":112251900,"identity":"c2994728-6ae6-4148-9a68-8128fd8b7b20","added_by":"auto","created_at":"2026-06-16 15:20:57","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":702901,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eHeatmap of associations between microbial species and diet.\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003ePanel A shows a heatmap of associations between microbial species and nutrients intake (nutrients\u003csub\u003e3 days\u003c/sub\u003e). Panel B shows a heatmap of associations between microbial species and food habits based on FFQ. Only species showing at least one association at q-value\u0026lt;0.25 are shown. Squares are colored according to the effect size of the association, as in the side legend. Asterisks mark level of significance (* q\u0026lt;0.25, **q\u0026lt;0.1, ***q\u0026lt;0.05).\u003c/p\u003e","description":"","filename":"12.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9888793/v1/a9e492a27ea50ddfbc5fdc0b.jpg"},{"id":112326940,"identity":"62abbba2-a5c0-4ab6-b06a-18c4638803f7","added_by":"auto","created_at":"2026-06-17 10:35:02","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":510429,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eAlpha and beta diversity by menstrual phase. \u003c/strong\u003e(A) Alpha diversity estimated using the Shannon index for each menstrual phase, showing overall diversity within samples for each menstrual phase. (B) Average beta diversity per women across menstrual phases. P-values shown in both panels are Bonferroni adjusted.\u003c/p\u003e","description":"","filename":"13.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9888793/v1/1b74cdf485aa757ab379966e.jpg"},{"id":112251903,"identity":"9d11dfeb-16b4-4863-8ba6-9655bfa68f16","added_by":"auto","created_at":"2026-06-16 15:20:57","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":386300,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eRelative abundance of dominant gut microbial taxa across weekly phases. \u003c/strong\u003eMicrobial species abundances are represented as proportions (%) in circular plots, stratified by phases.\u003c/p\u003e","description":"","filename":"14.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9888793/v1/3a06b3c70765486dc118ff6b.jpg"},{"id":112327059,"identity":"67f43c5a-ce11-4484-a049-62d58b668f87","added_by":"auto","created_at":"2026-06-17 10:35:46","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":312246,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eHeatmap of associations between microbial species, menstrual phase and reproductive hormones. \u003c/strong\u003ePanel A shows a heatmap of associations between microbial species and menstrual phase, using follicular phase as reference. Panel B shows a heatmap of associations between microbial species and reproductive hormones. Only species showing at least one association at q-value\u0026lt;0.25 are shown. Squares are colored according to the effect size of the association, as in the side legend. Asterisks mark level of significance (* q\u0026lt;0.25, **q\u0026lt;0.1, *** q\u0026lt;0.5).\u003c/p\u003e","description":"","filename":"15.jpg","url":"https://assets-eu.researchsquare.com/files/rs-9888793/v1/eb4839fb539be475922e3454.jpg"},{"id":112949493,"identity":"de666719-d0be-43ef-ac71-0d3def3064d6","added_by":"auto","created_at":"2026-06-24 14:22:15","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":640445,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9888793/v1/a725e3e7-5e9d-4b5f-b401-ac55f6c5749b.pdf"},{"id":112327254,"identity":"c41a62be-2077-435c-9a07-3afc58e5ed3e","added_by":"auto","created_at":"2026-06-17 10:36:29","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1694095,"visible":true,"origin":"","legend":"Supplementary Notes","description":"","filename":"SupplementaryNotesv2.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9888793/v1/1df434d1d5d063b537b0b79e.pdf"},{"id":112327273,"identity":"77ecd034-74d9-4922-a840-bd55d5b0eca7","added_by":"auto","created_at":"2026-06-17 10:36:31","extension":"xlsx","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":747734,"visible":true,"origin":"","legend":"Supplementary Tables","description":"","filename":"SupplementaryTablesv4.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-9888793/v1/d51f45cee81c40de7861e820.xlsx"},{"id":112251902,"identity":"9486e9aa-a0ad-4e63-835b-f2295db13b43","added_by":"auto","created_at":"2026-06-16 15:20:57","extension":"pdf","order_by":3,"title":"","display":"","copyAsset":false,"role":"supplement","size":3278179,"visible":true,"origin":"","legend":"Supplementary figures","description":"","filename":"SupplementaryFiguresv3.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9888793/v1/0dd543a49a8689262328b553.pdf"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Gut Microbiome Variation Across Menstrual Phases: Distinct Roles of Diet and Physiological Hormonal Fluctuations in Healthy Women","fulltext":[{"header":"Introduction","content":"\u003cp\u003eDiet is a dominant factor associated with interindividual variability in gut microbiome. Recent large-scale studies based on well-characterized cohorts have documented how dietary patterns shape microbiome composition and, in turn, associated the microbiome with metabolic biomarkers indicative of cardiometabolic health \u003csup\u003e\u003cspan additionalcitationids=\"CR2 CR3 CR4\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. For instance, Fackelmann and colleagues demonstrated that common dietary profiles (omnivore, vegetarian, and vegan) produce distinct gut microbiota signatures reflecting the predominant foods consumed in each diet. A diet with red meat consumption, for example, is marked by the presence of characteristic gut bacteria, such as \u003cem\u003eRuminococcus torques\u003c/em\u003e and \u003cem\u003eAlistipes putredinis\u003c/em\u003e, which are largely absent in vegan individuals \u003csup\u003e\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e\u003c/sup\u003e. Notably, even single foods can leave unique microbial \u0026ldquo;fingerprints\u0026rdquo;, as in the case of coffee consumption, which is strongly associated with the presence of \u003cem\u003eLawsonibacter asaccharolyticus\u003c/em\u003e \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWhile research on the diet-microbiome interaction has advanced considerably, the influence of other host factors remains yet to be understood. For example, several studies have shown that sex is a significant determinant of microbiota diversity in human populations \u003csup\u003e\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e,\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u003c/sup\u003e, a difference largely attributed to sex steroid hormones. Men and women in fact diverge in gut microbiome composition\u0026mdash;such as altered Firmicutes/Bacteroidetes ratios\u0026mdash;and these differences emerge through puberty (Kim 2020; Sisk-Hackworth 2023; \u0026Ouml;zkurt 2021) and are reduced after menopause \u003csup\u003e\u003cspan additionalcitationids=\"CR10\" citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eMoreover, gut bacteria are known metabolizers of steroid hormones. Several taxa express the beta-glucuronidase and beta-glucosidase enzymes, which hydrolyze conjugated (inactive) estrogens that have been excreted by the liver into the gut \u003csup\u003e\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u003c/sup\u003e. This process liberates active estrogens, enabling their reabsorption through enterohepatic circulation \u003csup\u003e\u003cspan additionalcitationids=\"CR14\" citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e. The collection of these microbes and their metabolic products capable of metabolizing estrogens have been termed the \u0026ldquo;estrobolome\u0026rdquo;. Both reduced gut microbial diversity and diminished microbial β‑glucuronidase activity observed in post-menopausal women, compared to pre-menopausal women, have been associated with systemic inflammation and cardiometabolic risk \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. Other complex mechanisms involving microbial enzymes have been proposed for the metabolization of other steroid hormones, such as progesterone, but current knowledge remains limited \u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eSex steroid hormones also play direct roles in digestive physiology. For example, estrogen contributes to gastrointestinal motility and mucosal health, and its decline during menopause has been associated with increased gastrointestinal symptoms \u003csup\u003e\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u003c/sup\u003e. Evidence from animal models further highlights the relevance of hormone-microbiome interactions: ovariectomized mice fed with a low-fat diet exhibit marked weight gain and increased inflammatory status, compared with high-fat-fed mice, suggesting that the metabolic impact of gut microbiome-diet interaction is contingent on female sex hormone status \u003csup\u003e\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eTo date, studies in human cohorts have typically examined either sex differences or dietary influences, but few have considered diet and hormones simultaneously. Studies focusing on women have been predominantly cross-sectional, frequently contrasting healthy and pathological states \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e or comparing pre- and post-menopausal women \u003csup\u003e\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e,\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e\u003c/sup\u003e. Furthermore, only small-scale studies have directly measured reproductive hormones along with gut microbiome \u003csup\u003e\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e,\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e\u003c/sup\u003e. Conversely, many longitudinal studies on diet and microbiota lack direct information on reproductive hormones hormones and hormonal status or rely solely on the information on the number of days since menstruation \u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eTo address these gaps, we propose a comprehensive study of gut microbiome profiles across the menstrual cycle in healthy Italian reproductive-age women from the Women4Health cohort \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e, incorporating detailed dietary assessments and direct measurements of reproductive hormones \u0026ndash; including not only estrogen and progesterone, but also hormones not yet explored for their interaction with gut microbiota, such as follicle-stimulating hormone (FSH), luteinizing hormone (LH), and prolactin (PRL) (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e). By disentangling the relative contributions of diet and endogenous hormonal fluctuations to microbiome variation, we elucidate mechanisms uniquely relevant to female microbial ecology and health\u0026mdash;with important implications for precision microbiome-based interventions in women.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThe figure describes the sampling scheme along the menstrual cycle and the data collected and analyzed in this work.\u003c/p\u003e"},{"header":"Methods","content":"\u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStudy cohort description\u003c/h2\u003e \u003cp\u003eThe Women4Health (W4H) study cohort is a longitudinal, population-based study designed to investigate the interplay between hormonal fluctuations, the microbiome, and metabolic health in women of reproductive age. The study is the result of a collaborative effort between the Institute of Genetics and Biomedical Research (IRGB) of the National Research Council (CNR), the IRCCS Burlo-Garofolo in Trieste, the Azienda Ospedaliero Universitario (AOU) of Cagliari, and the IRCCS Policlinico Sant\u0026rsquo;Orsola in Bologna, Italy. Recruitment is ongoing, with the aim of enrolling up to 300 healthy women aged 18\u0026ndash;45 years with regular menstrual cycles and no major metabolic, gastrointestinal, or gynecological conditions.\u003c/p\u003e \u003cp\u003eFull details on the study design, recruitment protocol, and inclusion criteria have been previously published \u003csup\u003e\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e\u003c/sup\u003e. Briefly, at enrollment and during four follow-up visits corresponding to specific phases of the menstrual cycle, detailed questionnaire data and multiple biological samples (stool, saliva, blood and vaginal swabs) were collected.\u003c/p\u003e \u003cp\u003eW4H data collection took place in three Italian cities (Trieste, Cagliari and Bologna) representing the North-East, South, and Central-North regions of the country, respectively. The collection began at different times across these centers: in Trieste, it started in February 2022, in Cagliari in 2024 and in Bologna in 2025.\u003c/p\u003e \u003cp\u003eAt enrollment, volunteers enrolled were asked to complete a questionnaire. This captured a wide range of information, including demographic and anthropometric data, a detailed medical and gynecological history, family history of disease, exposure to tobacco (direct or indirect), use of prebiotics and probiotics and dietary habits through a food frequency questionnaire (FFQ). During follow-up visits, scheduled to capture four phases of the menstrual cycle (follicular (F), ovulatory (O), early and late luteal phases (EL and LL)), they completed additional questionnaires that included, among many items, information on stool consistency (using the Bristol Stool Form Scale) and the use of any medications or supplements. In addition to the food frequency questionnaire (FFQ), participants also recorded their daily dietary intake using MyFitnessPal, a user-friendly smartphone application that offers an extensive food database. The app allows for accurate tracking of daily calorie intake as well as macronutrients and other dietary components, including protein, fat, cholesterol, sugar, and fiber. Daily diaries were submitted by volunteers at the end of the observation period.\u003c/p\u003e \u003cp\u003eFrom February 2023 to 1 August 2025, 161 participants were enrolled and returned for at least one visit. Among these, 98 were from Trieste, 38 from Bologna, and 25 from Cagliari. While the FFQ was available for all of them, only a subset of 149 compiled and returned the MyFitnessPal diary. Descriptive statistics of the 161 women are given in \u003cb\u003eSupplementary Table\u0026nbsp;1.\u003c/b\u003e\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eMeasurement of hormone levels and reclassification of menstrual phases\u003c/h3\u003e\n\u003cp\u003eBlood samples were collected via venipuncture using silica gel-coated Vacutainer tubes. After collection, blood was left undisturbed at room temperature for 1 hour to allow clotting. Subsequently, serum was obtained by centrifugation at 4,000 rpm for 5 minutes at room temperature. A volume of 0.5 mL of serum was then aliquoted into two sterile safe-lock polypropylene tubes and stored at \u0026minus;\u0026thinsp;20\u0026deg;C. An aliquot of 1.5 ml serum was frozen and later used to measure 5 reproductive hormones: 17β-estradiol (E2), luteinizing hormone (LH), follicle-stimulating hormone (FSH), progesterone (P4) and prolactin (PRL). During each woman\u0026rsquo;s first visit, we also measured testosterone and thyroid parameters (thyroid stimulating hormone and free thyroxine (free-T\u003csub\u003e4\u003c/sub\u003e)) to exclude endocrine disorders. Hormones were measured in three separate batches (one for each year of recruitment: 2023, 2024, 2025) on the ADVIA Centaur CP Immunoassay System instrument (Siemens Healthineers).\u003c/p\u003e \u003cp\u003eHormone data processing involved three steps. First, we log-transformed the hormone measurements. Second, we corrected hormone levels for the batch effect by matching the means of batch 1 and 2 hormone distributions with the reference batch 3. For each hormone, we calculated the difference in means between batch 1 or 2 with batch 3 and subtracted this difference from the values of batch 1 and 2. Third, we found that serum storage time in the freezer affected hormone levels. To account for this, we fitted a Linear Mixed Model (LMM) with a fixed effect for storage time and a random intercept for ID (subject-specific baseline shift):\u003c/p\u003e \u003cp\u003ehormone\u0026thinsp;~\u0026thinsp;storage_time + (1|ID),\u003c/p\u003e \u003cp\u003eusing \u003cem\u003elme4\u003c/em\u003e R package v. 1.1\u0026ndash;36 and took the residuals of this model by subtracting the predicted population average of the storage time effect on hormone levels from the real hormone level:\u003c/p\u003e \u003cp\u003epheno_real - predict(lmm_fit, re.form\u0026thinsp;=\u0026thinsp;NA)\u003c/p\u003e \u003cp\u003eBy study design, samples were collected at four visits scheduled according to the last day of menstruation and intended to capture the major hormonal phases of the menstrual cycle: follicular, ovulatory, early luteal, and late luteal. Since each woman exhibits a unique temporal pattern of reproductive hormones, days since last menstruation may not always represent an accurate proxy for menstrual phases. Therefore, we developed an algorithm to redefine phases in a more biologically meaningful way, using directly measured concentrations of LH. Details have been described in a previous work from our group \u003csup\u003e\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e. Briefly, the procedure first determines the visit with maximal LH concentration as the ovulatory point. Once the LH peak is located, a predefined mapping shifts the remaining visits backward or forward while preserving their temporal order relative to ovulation. This shift may map some visits outside the valid phase range (1\u0026ndash;4). Such boundary cases are resolved by assigning the visit to the closest phase. The final output provides, for every visit of every participant, a biologically informed phase label. When multiple visits map to the same phase, downstream analyses use the average value of all corresponding measurements (reproductive hormones, phenotypes, dietary parameters, or microbiome relative abundances). For some women a phase could not be established and thus they were excluded from analyses. While we recognize that the optimal way to establish phases would have been to have information on the day of the subsequent menstruation or echographic measurements, these were not available to us. Therefore, in this work we present comparisons of diet and microbiome data across menstrual phases being aware that these are not clinically evaluated but still provide the most robust approximation using directly measured concentrations of oestradiol, progesterone, FSH, and LH.\u003c/p\u003e\n\u003ch3\u003eFecal sample processing\u003c/h3\u003e\n\u003cp\u003eIn each center and at enrollment, volunteers received stool kits along with detailed instructions for self-collection. To ensure sample stability at room temperature, the OMNIgene-GUT (OM-200, DNA Genotek) collection system was used. After sample collection, volunteers returned their stool samples to the Obstetrics and Gynecology Clinic within 24 hours, where samples were stored at room temperature. After homogenization with a vortex mixer, 250 microL of stool were aliquoted in two 2.0 mL safe-lock polypropylene tubes and stored at -80\u0026deg;C Celsius. Microbial DNA was extracted using the PowerFecalPro DNA kit (QIAGEN) and subsequently stored at -30\u0026deg;C.\u003c/p\u003e \u003cp\u003eA total of 551 samples were collected from 161 volunteers, across four visits. Microbial DNA extracted from these samples were shipped in two separate batches at Prebiomics S.r.l. and sequenced using the NovaSeq 6000 System platform (Illumina), targeting a sequencing output of approximately 9 Gb of 150 bp paired-end reads per sample. The first batch included samples from the initial 75 participants from Trieste, totaling 264 samples across the four visits. A second batch followed, repeating 1 sample that failed in batch 1 and completing the dataset with 287 samples.\u003c/p\u003e\n\u003ch3\u003eProfiling microbiome composition and function\u003c/h3\u003e\n\u003cp\u003eThe metagenomes were first analyzed using the KneadData tool (v0.12.0) to process metagenomic reads in FASTQ format. The reads were trimmed to a PHRED quality score of 30, and Illumina adapters were removed. After trimming, the integrated Bowtie2 tool (v2.5.3) \u003csup\u003e26\u003c/sup\u003e within KneadData was employed to eliminate reads that aligned to the human genome. The total number of sequencing reads, remaining after quality filtering and removal of human-origin reads and DNA concentration calculated at the sequencing center, was used as a covariate, as this could influence the detection of specific microbial taxa.\u003c/p\u003e \u003cp\u003eThe taxonomic composition of 551 metagenomes from 161 volunteers across the three collection sites (Trieste\u0026thinsp;=\u0026thinsp;98, Bologna\u0026thinsp;=\u0026thinsp;38, Cagliari\u0026thinsp;=\u0026thinsp;25) was profiled using MetaPhlAn (v4.1.1) with the MetaPhlAn marker gene database (mpa_vJun23) \u003csup\u003e\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e. MetaPhlAn provided species-level abundance estimates, referencing the CHOCOPhlAn (mpa_CHOCOPhlAn_202403) database. Additionally, microbial biochemical pathways were profiled using the HUMAnN pipeline (v3.9) \u003csup\u003e28\u003c/sup\u003e, integrated with the DIAMOND alignment tool (v2.1.9), the UniRef90 protein database and the ChocoPhlAn pan-genome database. The analysis identified 12 phyla, 303 classes, 317 orders, 353 families, 1,043 genera, 1,740 species and 185 pathways. After quality control applied to metagenomes and assessment of metadata, 9 samples were excluded from subsequent analysis due to low microbial diversity, inconsistencies in reported menstrual cycle information, or discrepancies in cycle phase classification. Following these exclusions, 542 fecal samples from 160 women were retained (Trieste\u0026thinsp;=\u0026thinsp;98, Bologna\u0026thinsp;=\u0026thinsp;37, Cagliari\u0026thinsp;=\u0026thinsp;25). When applying the redefined menstrual cycle phase criteria (see above), 489 samples from 160 women were included in the analyses.\u003c/p\u003e\n\u003ch3\u003eProcessing of dietary information\u003c/h3\u003e\n\u003cp\u003eDietary intake was assessed using two complementary approaches: a food frequency questionnaire (FFQ) and daily dietary records. The FFQ was administered at enrollment and was designed to estimate habitual intake of selected food items over typical daily or weekly time frames. All 161 participants with quality controlled (QCed) microbiome data provided the FFQ at enrollment, while daily diet registers were completed by a subset of only 149 participants. Specifically, they recorded all foods and beverages consumed throughout the study period using the MyFitnessPal application, which allows detailed documentation of food items and portion sizes and automatically calculates total energy (kcal) and nutrient intakes, including carbohydrates (g), fats (g), proteins (g), sugars (g), fiber (g), cholesterol (mg), and sodium (mg). Daily diet records were carefully manually screened for potential errors by identifying foods with outliers in nutrient intakes. Detailed information on dietary data processing and quality control is given in the \u003cb\u003eSupplementary Notes\u003c/b\u003e.\u003c/p\u003e \u003cp\u003eTo evaluate the agreement between the two dietary assessment methods, daily total energy (kcal) and nutrient intakes were estimated from the FFQ-derived consumption frequencies (see \u003cb\u003eSupplementary Notes\u003c/b\u003e) and compared with the daily averages (across the entire observation period) obtained from the daily dietary records. Differences in mean intakes between the two methods were assessed using paired Wilcoxon signed-rank tests, while relative agreement was evaluated with Spearman rank correlation coefficients. To account for multiple comparisons, p-values were adjusted using the Bonferroni correction. All analyses were performed using custom R script available in our GitHub folder (see \u003cb\u003eData Availability Statement\u003c/b\u003e).\u003c/p\u003e \u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eDifferences in nutritional intake across the menstrual phases\u003c/h2\u003e \u003cp\u003eLongitudinal changes in nutrient intake along the menstrual cycle were analyzed using Linear Mixed Models (LMM) implemented with the \u003cem\u003elmer\u003c/em\u003e function from the \u003cem\u003elme4\u003c/em\u003e package (v1.1-36) with p-values calculated via \u003cem\u003elmerTest\u003c/em\u003e package (v3.2-1). For each nutrient, we considered the 3-days average prior to fecal sample collection as variables (nutrient\u003csub\u003e3days\u003c/sub\u003e). Temporal variation was modeled using hormonal phase (treated as numeric, continuous variable) as a fixed effect, and participant ID was included as a random intercept to account for repeated measures. Age, BMI, and recruiting centers were included as covariates to control individual- and center-level variability. Thus, for each nutrient we fit the following model:\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003enutrient\u003csub\u003e3days \u003c/sub\u003e ~ phase + age + BMI + center + (1|ID)\u003c/h3\u003e\n\u003cp\u003eMultiple testing across nutrients was corrected for using the Bonferroni method. Potential non-linear effects of time were assessed initially using natural splines implemented in the \u003cem\u003esplines\u003c/em\u003e package (v4.4.3). Spline models were compared with reduced models via likelihood ratio tests, and, if significant, a linear model was also fitted and compared with the spline model to determine whether the relationship was essentially linear or non-linear. Predicted marginal effects of the main temporal variable were obtained using the \u003cem\u003eeffects\u003c/em\u003e package in R (v4.2-4) and visualized as line plots to illustrate nutrient trajectories across hormonal phases.\u003c/p\u003e \u003cp\u003eIn addition to LMM, pairwise comparisons of nutrient intake between hormonal phases were conducted using paired Wilcoxon signed-rank tests to account for repeated measures within individuals. Analyses were performed on age- and BMI-adjusted residuals. We used the Bonferroni method to correct for multiple testing.\u003c/p\u003e\n\u003ch3\u003eStatistical analyses of microbiome composition and function\u003c/h3\u003e\n\u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eAlpha and beta diversity\u003c/h2\u003e \u003cp\u003eWe computed alpha diversity using the Shannon entropy (diversity) of microbial abundances.\u003c/p\u003e \u003cp\u003eWe also derived species-level beta diversity using the Euclidean distance applied to species relative abundances after Centered Log-Ratio (CLR) transformation using the function \u003cem\u003evegdist\u003c/em\u003e of \u003cem\u003evegan\u003c/em\u003e R package (version 2.6_8). As CLR transformation cannot be applied to zero values, zeros in the datasets were adjusted by adding half of the lowest non-zero value to each cell. In addition, CLR-transformed abundances were adjusted for technical covariates (DNA concentration, total number of post-QC reads, and sequencing batch number) before calculating beta diversity and for all downstream analyses, using the following model:\u003c/p\u003e \u003cp\u003etaxa_adj=residuals (taxa_CLR\u0026thinsp;~\u0026thinsp;DNA_concentration\u0026thinsp;+\u0026thinsp;Reads_postQC\u0026thinsp;+\u0026thinsp;batch)\u003c/p\u003e \u003cp\u003e.\u003c/p\u003e \u003cp\u003eTo further characterize between-samples variation, principal coordinates analysis (PCoA) was applied to the beta diversity dissimilarity matrix within each visit using the \u003cem\u003ecmdscale()\u003c/em\u003e function (R version 4.4.1). This approach projects samples into a reduced-dimensional Euclidean space preserving pairwise dissimilarities.\u003c/p\u003e \u003cp\u003eTo evaluate differences between phases we first calculated the average beta diversity per each woman within visits. Pairwise differences of alpha and average beta diversity were tested using Wilcoxon rank-sum tests for comparisons across centers, and using paired Wilcoxon signed-rank tests for menstrual phase comparisons. P-values were adjusted for multiple testing using Bonferroni correction.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eAnalysis of variance\u003c/h2\u003e \u003cp\u003eTo estimate the contribution of host factors (such as phases, hormone levels\u0026mdash;E2, P4, FSH, LH, PRL, and dietary factors \u0026mdash; nutrient\u003csub\u003e3days\u003c/sub\u003e, FFQ, and dairy/FFQ clusters) to microbial diversity, we used permutational multivariate analysis of variance (PERMANOVA) with 9999 permutations, as implemented in the \u003cem\u003eadonis2\u003c/em\u003e function in the \u003cem\u003evegan\u003c/em\u003e R package (v. 2.6_8).\u003c/p\u003e \u003cp\u003eFirst, we fitted univariate PERMANOVA models for each variable. Each model was adjusted for BMI, Bristol stool scale score, age and time of faecal sample collection, and repeated measures were accounted for by restricting permutations within individuals with the 'strata' option. To obtain effect sizes (R\u0026sup2;) for each variable, we used marginal tests (by = 'margin').\u003c/p\u003e \u003cp\u003eVariables showing nominal significance (P\u0026thinsp;\u0026lt;\u0026thinsp;0.05) in the univariate analyses were subsequently selected and included in a multivariate PERMANOVA model. This combined model incorporated all the selected variables simultaneously, adjusting for the same set of covariates (BMI, Bristol Stool Scale score, age and faecal sample collection time) and accounting for repeated sampling through subject-level stratification. Marginal effects were used to quantify the independent contribution of each variable while controlling for all the other variables in the model.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eLongitudinal changes of microbiome and microbiome association with diet, menstrual phases and hormones\u003c/h2\u003e \u003cp\u003eFor the downstream analysis, we focused on bacterial genera, species, and pathways with a mean relative abundance greater than 0.1% and present in at least 20% of samples. This led to the inclusion of 257 taxa (71 genera and 117 species) and 185 pathways. Like bacterial abundances, microbial pathways abundances were CLR-transformed and adjusted for technical covariates.\u003c/p\u003e \u003cp\u003eDifferences in adjusted microbiome composition (microbial taxa abundances and pathways) between the three collection centers were evaluated using the Kruskal-Wallis test.\u003c/p\u003e \u003cp\u003eTo evaluate the impact of macronutrient intake on gut microbiome, we used the average intake estimated from daily diaries across 3 days prior to stool collection (nutrient\u003csub\u003e3days\u003c/sub\u003e). LMMs were applied for each macronutrient, adjusting for phase, age, body mass index (BMI), Bristol stool scale, recruiting center, and collection time (in the morning or during the day), with women ID included as a random effect.\u003c/p\u003e \u003cp\u003etaxa_adj\u0026thinsp;~\u0026thinsp;phase +nutrient\u003csub\u003e3days\u003c/sub\u003e+ Age\u0026thinsp;+\u0026thinsp;BMI\u0026thinsp;+\u0026thinsp;Bristol_stool_scale\u0026thinsp;+\u0026thinsp;Center\u0026thinsp;+\u0026thinsp;Collection_when + (1 | ID)\u003c/p\u003e \u003cp\u003eIn addition to nutrient intake, associations between the gut microbiome and dietary patterns were also assessed using nutrients derived from the FFQ (see \u003cb\u003eSupplementary Notes\u003c/b\u003e) and diet quality scores. Furthermore, dietary patterns, identified through principal component analysis and K-means clustering of macronutrient intake proportions, were also tested. Specifically, women were grouped into clusters based on similarities in their dietary profiles (see \u003cb\u003eSupplementary Notes\u003c/b\u003e). Association analysis with dietary patterns was performed using LMMs with the same set of covariates described above.\u003c/p\u003e \u003cp\u003eWe investigated first associations between microbial species and menstrual phases using paired Wilcoxon signed-rank tests on CLR abundances, after adjustment for technical covariates. Then, we also carried out this analysis using LMMs to account for the longitudinal structure of the data.\u003c/p\u003e \u003cp\u003eThe model included menstrual phase (treated as a categorical variable with follicular phase as the reference), age, BMI, Bristol stool scale, recruiting center, and time of collection as fixed effects, with women ID as a random intercept. The formula used was:\u003c/p\u003e \u003cp\u003etaxa_adj\u0026thinsp;~\u0026thinsp;phase\u0026thinsp;+\u0026thinsp;Age\u0026thinsp;+\u0026thinsp;BMI\u0026thinsp;+\u0026thinsp;Bristol_stool_scale\u0026thinsp;+\u0026thinsp;Center\u0026thinsp;+\u0026thinsp;Collection_when + (1 | ID)\u003c/p\u003e \u003cp\u003eTo investigate associations with reproductive hormones, we used a similar model, adding in turn each of the hormones to the model, and treating phase as numeric, continuous variable, as follows:\u003c/p\u003e \u003cp\u003etaxa_adj\u0026thinsp;~\u0026thinsp;phase\u0026thinsp;+\u0026thinsp;hormone\u0026thinsp;+\u0026thinsp;Age\u0026thinsp;+\u0026thinsp;BMI\u0026thinsp;+\u0026thinsp;Bristol_stool_scale\u0026thinsp;+\u0026thinsp;Center\u0026thinsp;+\u0026thinsp;Collection_when + (1 | ID)\u003c/p\u003e \u003cp\u003eFor the association with dietary features, hormonal phases and sex hormones, we considered significant the results with p-values\u0026thinsp;\u0026lt;\u0026thinsp;0.05 and Benjamini\u0026ndash;Hochberg\u0026ndash;corrected q\u0026thinsp;\u0026lt;\u0026thinsp;0.25.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec14\" class=\"Section2\"\u003e \u003ch2\u003eMediation analyses of hormones, microbiome and diet\u003c/h2\u003e \u003cp\u003eIn order to investigate if hormones mediate the impact of diet on the gut microbiome, we first identified microbial species and pathways that were significantly associated with both hormone levels and diet. Only those species/pathways showing associations with both factors were considered for mediation testing, as these factors can be intermediaries through which diet and hormones influence microbial composition. Mediation analyses were performed using the \u003cem\u003emediation\u003c/em\u003e R package (v.4.5.1) and applying a two-model approach. First, a mediator model was fitted using linear regression, in which hormone levels were modeled as a function of dietary exposure and covariates. Secondly, an outcome model was fitted, modelling the microbiome as a function of dietary exposure and the mediator. Both models were LMMs to account for repeated measures and included the same covariates as in the analyses described above. The mediation analysis was then conducted using the \u003cem\u003emediate()\u003c/em\u003e function to estimate the average causal mediation effect (ACME), representing the indirect effect through the hormone mediator; the average direct effect (ADE) of diet on the microbiome, independent of the mediator, and the total effect. Statistical significance and confidence intervals were estimated using non-parametric bootstrapping with 1,000 simulations.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eAssessment of enzymes involved in estrobolome activity\u003c/h2\u003e \u003cp\u003eWe quantified 7,976 bacterial enzyme abundances mapping the HUMAnN output to level 4 enzyme commission (EC) categories. According to HUMAnN protocol, we normalized counts by total number of postQC sequencing reads and derived relative abundances. They were CLR transformed and adjusted for technical covariates. Among those with prevalence\u0026thinsp;\u0026gt;\u0026thinsp;20%, we selected enzymes that were potentially involved in the \u003cem\u003eestrobolome\u003c/em\u003e, in particular enzymes involved in glycoconjugate metabolism (beta-glucuronidase, beta-galactosidase, beta-glucosidase, 6-phospho-beta-glucosidase) and sulfate metabolism (arylsulfatase type 1 and aryl-sulfate sulfotransferase) \u003csup\u003e\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e,\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e,\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e\u003c/sup\u003e, and evaluated the association between their abundances, menstrual phases, and circulating reproductive hormones levels using paired Wilcoxon signed-rank tests. We also repeated the association using LMMs; however, outliers (defined as samples with enzyme abundance\u0026thinsp;\u0026gt;\u0026thinsp;3 standard deviation from the mean) were removed prior to model fitting. Specifically, we used the following LMM models:\u003c/p\u003e \u003cp\u003eenzyme_adj\u0026thinsp;~\u0026thinsp;phase\u0026thinsp;+\u0026thinsp;Age\u0026thinsp;+\u0026thinsp;BMI\u0026thinsp;+\u0026thinsp;Bristol_stool_scale\u0026thinsp;+\u0026thinsp;Center+ Collection_when + (1 | ID)\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eand\u003c/h2\u003e \u003cp\u003eenzyme_adj\u0026thinsp;~\u0026thinsp;phase\u0026thinsp;+\u0026thinsp;hormone\u0026thinsp;+\u0026thinsp;Age\u0026thinsp;+\u0026thinsp;BMI\u0026thinsp;+\u0026thinsp;Bristol_stool_scale\u0026thinsp;+\u0026thinsp;Center\u0026thinsp;+\u0026thinsp;Collection_when + (1 | ID)\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eWithin the Wome4Health cohort we analysed reproductive hormones, gut microbiome and dietary information across 4 visits during the menstrual cycle for 160 women not taking oral contraceptives. Participants were healthy women with a mean BMI of 21.93 (S.D. 2.9) and mean age of 29.2 years old (S.D. 2.9 years). Detailed sample characteristics are provided in \u003cb\u003eSupplementary Table \u003cspan refid=\"MOESM1\" class=\"InternalRef\"\u003eS1\u003c/span\u003e\u003c/b\u003e.\u003c/p\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eConsistency between dietary patterns and nutrients intake from FFQ and daily records\u003c/h2\u003e \u003cp\u003e To investigate the impact of diet on microbiome, we first evaluated consistency between two types of dietary information collected in the study: a frequency food questionnaire (FFQ) administered at enrollment, and a daily diet registry compiled by participants on the MyFitnessPal app. The latest was available only for a subset of 149 out of the 161 women. From the FFQ, we estimated the average macronutrients per day using standard portions obtained from the Italian Society of Human Nutrition\u0026rsquo;s table (SINU \u0026ndash; Societ\u0026agrave; Italiana di Nutrizione Umana, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://eng.sinu.it/\u003c/span\u003e\u003cspan address=\"https://eng.sinu.it/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003cspan type=\"Underline\" class=\"Underline\" name=\"Emphasis\"\u003e)\u003c/span\u003e and nutrient data from the Italian Food Composition Database (BDA) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://bda.ieo.it/\u003c/span\u003e\u003cspan address=\"https://bda.ieo.it/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e ) (see \u003cb\u003eSupplementary Notes\u003c/b\u003e). Daily diaries underwent strict quality control to identify and correct errors in nutrients definition (\u003cb\u003eSupplementary Notes\u003c/b\u003e). We observed that the average nutrient intake calculated from all daily diaries collected across the entire study period differed significantly from the estimated average macronutrient intake derived from FFQs (\u003cb\u003eSupplementary Table\u0026nbsp;2\u003c/b\u003e). Daily diaries reported, on average, significantly higher intakes of total energy, carbohydrates, fats, proteins and sugars compared with FFQs (Bonferroni-adjusted p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.001 for all comparisons). In contrast, FFQs estimated significantly higher intakes of cholesterol and fiber than daily diaries (Bonferroni-adjusted p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.001). Sodium intake did not differ significantly between the two dietary assessment methods after correction for multiple comparisons (Bonferroni-adjusted p-value\u0026thinsp;\u0026gt;\u0026thinsp;0.05).\u003c/p\u003e \u003cp\u003eNevertheless, measurements were highly correlated (\u003cb\u003eSupplementary Table\u0026nbsp;2\u003c/b\u003e), except for sugar and sodium, probably reflecting the difficulties in estimating sweets/snacks on FFQs.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec19\" class=\"Section2\"\u003e \u003ch2\u003eAssociation between diet and gut microbiome species abundance\u003c/h2\u003e \u003cp\u003eSince our data was collected along a menstrual cycle, we first evaluated the effect of the hormonal phase on each nutrient intake (using the average of the 3-day data recorded prior to fecal sample collection, averaged within the same hormonal phases, nutrients\u003csub\u003e3days\u003c/sub\u003e, see \u003cb\u003eMethods\u003c/b\u003e) (\u003cb\u003eSupplementary Table\u0026nbsp;3\u003c/b\u003e). While some associations showed a nominal p-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05, none remained significant after Bonferroni correction for multiple comparisons, indicating that diet patterns are not significantly different across the menstrual cycle.\u003c/p\u003e \u003cp\u003eWe then examined how nutrients intake influenced gut microbiome species and pathways. We characterized the taxonomic and functional composition of the gut microbiome of 542 fecal samples collected from 160 women across the three centers along four menstrual phases. Shotgun metagenomic data identified, within the Bacteria kingdom, a total of 12 phyla, 303 classes, 317 orders, 353 families, 1,043 genera, 1,740 species, and 532 metabolic pathways (\u003cb\u003eSupplementary Fig.\u0026nbsp;1\u003c/b\u003e). Given our interest in fine-scale microbial variation and metabolic pathways, subsequent analyses focused primarily on lower taxonomic ranks, particularly at the species level and on microbial pathways. To reduce sparsity and increase robustness of our data, we applied stringent filtering criteria retaining only features with a relative abundance\u0026thinsp;\u0026gt;\u0026thinsp;0.1% and present in at least 20% of samples. After filtering, 117 species and 185 pathways were retained for downstream analyses. We identified 26 associations of nutrients\u003csub\u003e3days\u003c/sub\u003e with species abundances and 28 with microbial pathways at BH q-value\u0026thinsp;\u0026lt;\u0026thinsp;0.25 (\u003cb\u003eSupplementary Tables\u0026nbsp;4 and 5\u003c/b\u003e, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA, \u003cb\u003eSupplementary Fig.\u0026nbsp;2\u003c/b\u003e). Among these, we see some expected associations. For example, higher fiber intake was associated with increased abundance of \u003cem\u003eRoseburia hominis\u003c/em\u003e (effect estimate\u0026thinsp;=\u0026thinsp;0.215, q\u0026thinsp;=\u0026thinsp;0.112, where the estimates represent changes in standard deviation units of CLR abundance), a butyrate‑producing taxon that has been previously linked to beneficial metabolic effects and enrichment with dietary fiber consumption \u003csup\u003e\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e. Likewise, a negative association between sugar intake and abundance of \u003cem\u003eClostridium sp. AM22_11AC\u003c/em\u003e (effect estimate\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;0.155, q\u0026thinsp;=\u0026thinsp;0.060) is consistent with prior studies indicating that diets higher in simple sugars can alter gut microbiota composition, including reductions in certain Firmicutes‑associated taxa, such as \u003cem\u003eClostridium\u003c/em\u003e \u003csup\u003e\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e\u003c/sup\u003e. Additionally, we also found that fat intake was positively associated with \u003cem\u003eBifidobacterium bifidum\u003c/em\u003e (effect estimate\u0026thinsp;=\u0026thinsp;0.101, q\u0026thinsp;=\u0026thinsp;0.158), a species often investigated in studies related to diet. Although we observed a positive trend between fat intake and \u003cem\u003eBifidobacterium bifidum\u003c/em\u003e abundance, evidence from human studies indicates that the effects of dietary fat on \u003cem\u003eBifidobacterium spp.\u003c/em\u003e are highly context‑dependent and influenced by both the quantity and type of fat consumed \u003csup\u003e\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e\u003c/sup\u003e. Interestingly, we observed that protein intake also showed a modest negative association with \u003cem\u003eAcidaminococcus intestini\u003c/em\u003e (effect estimate\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;0.093, q\u0026thinsp;=\u0026thinsp;0.101). This bacterium is known for fermenting amino acids, such as glutamate, linking to potential roles in protein and amino acid metabolism in the gut \u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e. Among microbial pathways, the strongest association was observed between higher cholesterol intake and reduced abundance of pathways involved in bacterial lipid biosynthesis, including anaerobic gondoate biosynthesis (effect estimate\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;0.045, q\u0026thinsp;=\u0026thinsp;0.020) and cis-vaccenate biosynthesis (effect estimate\u0026thinsp;=\u0026thinsp;\u0026minus;\u0026thinsp;0.043, q\u0026thinsp;=\u0026thinsp;0.023).\u003c/p\u003e \u003cp\u003eFurthermore, using the same mixed-effects modeling framework applied in previous association analyses, we investigated the relationship between gut microbial species abundance and habitual food consumption based on FFQ data (weekly intake frequencies). We identified several diet\u0026ndash;microbiome associations (54 with species and 144 with pathways), with beverage-related patterns showing the strongest signals (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eB, \u003cb\u003eSupplementary Fig.\u0026nbsp;3, Supplementary Tables\u0026nbsp;6 and 7\u003c/b\u003e). In particular, coffee consumption was strongly positively associated with a butyrate-producing \u003cem\u003eLawsonibacter asaccharolyticus\u003c/em\u003e (effect estimate\u0026thinsp;=\u0026thinsp;0.464, q\u0026thinsp;=\u0026thinsp;1.38\u0026times;10⁻⁵), confirming previous large-scale population-based metagenomic studies \u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. Interestingly, \u003cem\u003eLawsonibacter asaccharolyticus\u003c/em\u003e also showed a positive association with yogurt consumption (effect estimate\u0026thinsp;=\u0026thinsp;0.256, q\u0026thinsp;=\u0026thinsp;0.063). This association was not significantly altered (effect estimate\u0026thinsp;=\u0026thinsp;0.21, q\u0026thinsp;=\u0026thinsp;0.10) after adjustment for coffee intake, indicating an independent relationship with yogurt consumption. This specific taxon has not been previously described among the microbiome shifts associated with yogurt intake. Additional positive associations with coffee consumption were observed for \u003cem\u003eAlistipes communis\u003c/em\u003e and \u003cem\u003eAlistipes ihumii\u003c/em\u003e, both members of the genus \u003cem\u003eAlistipes\u003c/em\u003e which has been previously linked to higher caffeine consumption \u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eAnother known association was observed between yogurt intake and \u003cem\u003eStreptococcus thermophilus\u003c/em\u003e (effect estimate\u0026thinsp;=\u0026thinsp;0.344, q\u0026thinsp;=\u0026thinsp;0.0068), a well-established yogurt starter species \u003csup\u003e\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e. Among the pathways, particularly intriguing was the association between olive oil consumption and the superpathway of beta-D-glucuronosides degradation (effect estimate\u0026thinsp;=\u0026thinsp;0.106, q\u0026thinsp;=\u0026thinsp;0.014), a complex microbial pathway that includes among key enzymes the beta-glucuronidase activity. This may suggest a potential link between diet, microbial deconjugation capacity, and hormone-related metabolism.\u003c/p\u003e \u003cp\u003eSeveral other associations were found at a lower level of significance, but larger studies are needed to confirm their validity.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003ePanel A shows a heatmap of associations between microbial species and nutrients intake (nutrients\u003csub\u003e3 days\u003c/sub\u003e). Panel B shows a heatmap of associations between microbial species and food habits based on FFQ. Only species showing at least one association at q-value\u0026thinsp;\u0026lt;\u0026thinsp;0.25 are shown. Squares are colored according to the effect size of the association, as in the side legend. Asterisks mark level of significance (* q\u0026thinsp;\u0026lt;\u0026thinsp;0.25, **q\u0026thinsp;\u0026lt;\u0026thinsp;0.1, ***q\u0026thinsp;\u0026lt;\u0026thinsp;0.05).\u003c/p\u003e \u003cp\u003eFinally, we used several metrics of dietary patterns, specifically dietary definitions derived by clustering the observed macronutrients intake, and established diet quality indices (the Alternate Mediterranean Diet (aMED), the Healthy Eating Index (HEI) and Alternate Healthy Eating Index (aHEI), and the Dietary Approaches to Stop Hypertension (DASH score) and associated these with species abundances, alpha and beta diversity (\u003cb\u003eSupplementary Table\u0026nbsp;8, Supplementary Table\u0026nbsp;9, Supplementary Table\u0026nbsp;10\u003c/b\u003e). Alpha and beta diversity were not significantly associated with diet scores. Therefore, the overall diet habits were overall homogeneous in our data and did not drive major changes of the gut microbiome. Only a few associations were observed at the species level.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec20\" class=\"Section2\"\u003e \u003ch2\u003eContribution of diet, hormone levels and menstrual phase to microbiome\u003c/h2\u003e \u003cp\u003eTo quantify the relative influence of diet, menstrual phases and reproductive hormones on gut microbiome, we performed a series of Permutational Multivariate Analysis of Variance (PERMANOVA) analyses on species-level beta diversity (see \u003cb\u003eMethods\u003c/b\u003e). We modeled beta diversity as a function of menstrual phase, sex hormones, dietary patterns derived from FFQ and short-term food diary clusters. We first performed univariate PERMANOVA models to evaluate contribution of each variable while adjusting for covariates. Several dietary components and hormonal variables showed nominal associations with microbial beta diversity, although all effect sizes were small. Among dietary factors, wine (R\u0026sup2; = 0.58%, p\u0026thinsp;=\u0026thinsp;0.018), fish (R\u0026sup2; = 0.57%, p\u0026thinsp;=\u0026thinsp;0.027), olive oil (R\u0026sup2; = 0.65%, p\u0026thinsp;=\u0026thinsp;0.026), butter consumption (R\u0026sup2; = 0.54%, p\u0026thinsp;=\u0026thinsp;0.032), and fiber intake from daily diaries (R\u0026sup2; = 0.47%, p\u0026thinsp;=\u0026thinsp;0.016) reached nominal significance. Among hormonal variables, P4 (R\u0026sup2; = 0.15%, p\u0026thinsp;=\u0026thinsp;0.024), LH (R\u0026sup2; = 0.21%, p\u0026thinsp;=\u0026thinsp;0.004) and FSH (R\u0026sup2; = 0.31%, p\u0026thinsp;=\u0026thinsp;0.005) were also nominally associated with beta diversity. In contrast, menstrual phase and dietary pattern clusters derived from FFQ and short-term diaries showed no association \u003cb\u003e(Supplementary Table\u0026nbsp;11)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003eVariables with nominal significance in univariate analyses were included in a multivariable PERMANOVA model. The full model explained 7.6% of the total variance in microbial beta diversity (p\u0026thinsp;=\u0026thinsp;0.001), meaning that a large fraction of the variability remains unexplained. When estimating the independent contribution of each variable within the multivariable model using marginal tests, only fiber intake (R\u0026sup2; = 0.46%, p\u0026thinsp;=\u0026thinsp;0.017) and Bristol Stool Scale score (R\u0026sup2; = 0.69%, p\u0026thinsp;=\u0026thinsp;0.016) remained statistically significant. All other variables, such as dietary components (wine, fish, olive oil, butter) and hormones (P4, LH, FSH), were no longer significant \u003cb\u003e(Supplementary Table\u0026nbsp;11)\u003c/b\u003e.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec21\" class=\"Section2\"\u003e \u003ch2\u003eCharacterization of gut microbiome across menstrual phases\u003c/h2\u003e \u003cp\u003eNext, we focused on changes in gut microbiome during the menstrual cycle. Intriguingly, alpha diversity showed significant changes, in particular it was lower in the follicular phase compared to ovulatory phase (Bonferroni-adjusted Wilcoxon test p-value\u0026thinsp;=\u0026thinsp;0.038), while other phase-to-phase comparisons were not statistically significant (all nominal p\u0026thinsp;\u0026gt;\u0026thinsp;0.05) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA, \u003cb\u003eSupplementary Table\u0026nbsp;12\u003c/b\u003e). Likewise, average beta diversity per woman significantly increased from earlier phases to late luteal phase (follicular versus late luteal, Bonferroni-adjusted paired Wilcoxon test p\u0026thinsp;=\u0026thinsp;8.7 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;6\u003c/sup\u003e; ovulatory versus late luteal, p\u0026thinsp;=\u0026thinsp;6.4 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;4\u003c/sup\u003e; early luteal versus late luteal, p\u0026thinsp;=\u0026thinsp;4.2 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;6\u003c/sup\u003e) (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB, \u003cb\u003eSupplementary Table\u0026nbsp;12\u003c/b\u003e). These results suggested that microbial community composition undergoes a reorganization during the menstrual cycle. We also examined the association between individual hormones and beta diversity. Spearman correlation analysis between PCoA1 of beta diversity and hormones revealed a significant positive association with FSH during the early luteal phase (Spearman correlation ρ\u0026thinsp;=\u0026thinsp;0.20, p\u0026thinsp;=\u0026thinsp;0.005, see \u003cb\u003eSupplementary Fig.\u0026nbsp;4\u003c/b\u003e), while no significant correlations were observed in other phases or for LH, P4, PRL, and E2 .\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eMicrobiome composition was overall stable across visits. At the genus level, the highest abundant bacterial genera were \u003cem\u003eBacteroides\u003c/em\u003e and \u003cem\u003ePhocaeicola\u003c/em\u003e whereas, at species level, microbiome was mostly dominated by \u003cem\u003eFaecalibacterium prausnitzii, Bacteroides uniformis\u003c/em\u003e and \u003cem\u003ePhocaeicola vulgatus\u003c/em\u003e (\u003cb\u003eSupplementary Fig.\u0026nbsp;5\u003c/b\u003e; \u003cb\u003eSupplementary Fig.\u0026nbsp;6\u003c/b\u003e). Stratification by menstrual phases revealed only minor differences in single species abundances \u003cb\u003e(\u003c/b\u003eFig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003e\u003cb\u003e)\u003c/b\u003e. Slight differences at the level of species abundances were instead seen between recruiting centers (\u003cb\u003eSupplementary Fig.\u0026nbsp;7\u003c/b\u003e), and thus, the city of collection was used as covariate in all subsequent analyses.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eNext, we investigated associations between gut microbial species abundance and pathways with menstrual phases using first Wilcoxon signed-rank tests and then linear mixed-effects models. In the paired Wilcoxon signed-rank analysis, \u003cem\u003eBacteroides faecis\u003c/em\u003e was the only significant species (q‑value\u0026thinsp;=\u0026thinsp;0.21), showing a difference between the follicular and early luteal phases (\u003cb\u003eSupplementary Table\u0026nbsp;13\u003c/b\u003e). In the linear mixed-effects models, in addition to \u003cem\u003eB. faecis\u003c/em\u003e, we detected other 10 species and 5 pathways exhibiting differences in abundance between the follicular and subsequent phases (\u003cb\u003eSupplementary Table\u0026nbsp;14\u003c/b\u003e; \u003cb\u003eSupplementary Table\u0026nbsp;15)\u003c/b\u003e. \u003cem\u003eBlautia massiliensis\u003c/em\u003e showed increased abundance in ovulatory (effect estimate\u0026thinsp;=\u0026thinsp;0.20, q\u0026thinsp;=\u0026thinsp;0.18), early luteal (effect estimate\u0026thinsp;=\u0026thinsp;0.28, q\u0026thinsp;=\u0026thinsp;0.03) and late luteal phase (effect estimate\u0026thinsp;=\u0026thinsp;0.26, q\u0026thinsp;=\u0026thinsp;0.15) as compared to the follicular phase. Conversely, \u003cem\u003eLacrimispora amygdalina\u003c/em\u003e decreased in late luteal (effect estimate = -0.45, q\u0026thinsp;=\u0026thinsp;0.04), while \u003cem\u003eRoseburia intestinalis\u003c/em\u003e decreased in early and late luteal phases (effect estimate = -0.39, q\u0026thinsp;=\u0026thinsp;0.17; and effect estimate = -0.41, q\u0026thinsp;=\u0026thinsp;0.23, respectively). Other notable associations included increased abundance of \u003cem\u003eDysosmobacter welbionis\u003c/em\u003e (late luteal, effect estimate\u0026thinsp;=\u0026thinsp;0.28, q\u0026thinsp;=\u0026thinsp;0.17), \u003cem\u003eClostridium spp\u003c/em\u003e. (early luteal, \u003cem\u003eClostridium sp AF36 4\u003c/em\u003e, effect\u0026thinsp;=\u0026thinsp;0.23, q\u0026thinsp;=\u0026thinsp;0.22; \u003cem\u003eClostridium sp AM22 11AC\u003c/em\u003e, effect estimate\u0026thinsp;=\u0026thinsp;0.19, q\u0026thinsp;=\u0026thinsp;0.23), and \u003cem\u003eRoseburia hominis\u003c/em\u003e (late luteal, effect estimate\u0026thinsp;=\u0026thinsp;0.42, q\u0026thinsp;=\u0026thinsp;0.24) (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eA\u003cb\u003e).\u003c/b\u003e\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec22\" class=\"Section2\"\u003e \u003ch2\u003eRelationship between serum levels of reproductive hormones hormones and the gut microbiome\u003c/h2\u003e \u003cp\u003eNext, we evaluated the relationship between gut microbiome species and pathways abundance with the five circulating reproductive hormone levels: progesterone (P4), luteinizing hormone (LH), follicle-stimulating hormone (FSH), 17β-estradiol (E2) and prolactin (PRL), using linear mixed-effects models. A total of 28 species and 22 pathways showed significant associations with reproductive hormones (\u003cb\u003eSupplementary Table\u0026nbsp;16; Supplementary Table\u0026nbsp;17;\u003c/b\u003e Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003eB). Interestingly, we observed the highest number of significant associations not for the most well-studied hormones such as 17β-estradiol (N\u0026thinsp;=\u0026thinsp;3) and progesterone (N\u0026thinsp;=\u0026thinsp;8), but rather for FSH and LH (N\u0026thinsp;=\u0026thinsp;20 and 19, respectively). For example, LH and FSH were positively associated with \u003cem\u003ePhocaeicola vulgatus\u003c/em\u003e (also known as \u003cem\u003eBacteroides vulgatus\u003c/em\u003e) (q\u0026thinsp;=\u0026thinsp;0.06 and q\u0026thinsp;=\u0026thinsp;0.05, respectively), a bacterial species known to induce PCOS-like phenotype in mice \u003csup\u003e\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e. Thus, while the association with these hormones is intriguing, the effect directions are not always consistent with the physiology of PCOS patients, characterized by high LH and low-to-normal FSH. Furthermore, these hormones were also positively associated with another species of the same genus \u003cem\u003ePhocaeicola massiliensis\u003c/em\u003e (q\u0026thinsp;=\u0026thinsp;0.08 and q\u0026thinsp;=\u0026thinsp;0.08, respectively), for which no evidence exists, to our knowledge, for an involvement in gynecological or endocrine disorders, and with \u003cem\u003eFaecalibacterium prausnitzii\u003c/em\u003e (q\u0026thinsp;=\u0026thinsp;0.23 and q\u0026thinsp;=\u0026thinsp;0.06, respectively), for which association with PCOS is ambiguous \u003csup\u003e\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e,\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e. Associations with 17β-estradiol highlighted a negative link with \u003cem\u003eAkkermansia muciniphila\u003c/em\u003e (q\u0026thinsp;=\u0026thinsp;0.22), albeit this is in contrast with studies carried out in vitro showing an increase of this species after estradiol supplementation \u003csup\u003e\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e\u003c/sup\u003e. Notably, the effect size of all the significant associations between menstrual phases, reproductive hormones, bacterial species and pathways, remained consistent even when adjusting the model with food habits from FFQs or nutrients\u003csub\u003e3days\u003c/sub\u003e from daily diet (\u003cb\u003eSupplementary Fig.\u0026nbsp;8\u003c/b\u003e).\u003c/p\u003e \u003cdiv id=\"Sec23\" class=\"Section3\"\u003e \u003ch2\u003eDiet directly shapes gut microbiome taxa and pathway abundance independent of hormonal mediation\u003c/h2\u003e \u003cp\u003eTo investigate whether hormonal fluctuations mediate the impact of diet on the gut microbiome, we considered species that were significantly associated with both circulating hormone levels and short-term dietary intake (nutrients\u003csub\u003e3days\u003c/sub\u003e), as these taxa represent plausible intermediates through which diet and hormones may influence microbial composition. Five bacterial species were included in the mediation analysis: \u003cem\u003eAlistipes shahii\u003c/em\u003e (linked to prolactin and dietary cholesterol), \u003cem\u003eBifidobacterium bifidum\u003c/em\u003e (linked to LH, FSH, and dietary fat), \u003cem\u003eGGB9758_SGB15368\u003c/em\u003e (linked to FSH and dietary fat), \u003cem\u003eMediterraneibacter faecis\u003c/em\u003e (linked to LH and dietary sugar), and \u003cem\u003ePhascolarctobacterium faecium\u003c/em\u003e (linked to FSH and dietary sugar). We also applied the same approach to microbial functional pathways to explore any potential mediation by hormones. The pathway PWY0.1479 (tRNA processing) was associated with FSH and LH, as well as with dietary intake of calories, fiber, and proteins, and was therefore included in the mediation analysis.\u003c/p\u003e \u003cp\u003eThe average causal mediation effect (ACME) was not statistically significant (all p-values\u0026thinsp;\u0026gt;\u0026thinsp;0.33), indicating that in our data hormones did not mediate the association between diet and bacterial abundance. The average direct effect (ADE) of diet on bacterial abundance remained significant for most bacteria (e.g., \u003cem\u003eB. bifidum\u003c/em\u003e, ADE p-value\u0026thinsp;=\u0026thinsp;0.004\u0026ndash;0.01; \u003cem\u003eM. faecis\u003c/em\u003e, ADE p-value\u0026thinsp;=\u0026thinsp;0.018; \u003cem\u003eP. faecium\u003c/em\u003e, ADE p-value\u0026thinsp;=\u0026thinsp;0.002), and the total effect was also significant. The proportion of the total effect mediated through hormones was non-significant (\u003cb\u003eSupplementary Table\u0026nbsp;18\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eLikewise, for the tested pathway (PWY0-1479: tRNA processing), the ACME was not statistically significant, with confidence intervals consistently crossing zero. In contrast, the ADE of diet on pathway abundance remained significant for multiple exposures, such as protein, calorie, and fiber intake (ADE p-value\u0026thinsp;=\u0026thinsp;0.004\u0026ndash;0.016), and the total effect of diet on pathway was also significant. The proportion of the total effect mediated via hormones was also modest (ranging from ~\u0026thinsp;0% to \u0026minus;\u0026thinsp;0.3%) and non-significant (\u003cb\u003eSupplementary Table\u0026nbsp;19\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eThese findings indicated that the associations between short-term dietary intake and both gut bacterial abundances and metabolic pathways are predominantly direct, rather than mediated by hormonal fluctuations. This supports the concept in which diet directly shapes microbial composition and functional capacity, independent of circulating hormone levels.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec24\" class=\"Section2\"\u003e \u003ch2\u003eEstrobolome enzyme abundance across menstrual phases\u003c/h2\u003e \u003cp\u003eMany of the bacteria we found to be associated with hormones, possess genes encoding the \u0026ldquo;estrobolome\u0026rdquo; - the collection of microbial enzymes hypothesized to participate in human estrogen metabolism, with a potential for progesterone metabolism. Therefore, we thought to specifically evaluate the abundance of six microbiome enzymes known or hypothesized to be involved in the estrobolome activity, including established enzymes such as beta-glucuronidase and beta-galactosidase, and several potentially relevant enzymes: arylsulfatase, aryl-sulfate sulfotransferase, 6-phospho-beta-glucosidase and beta-glucosidase (\u003cb\u003eSupplementary Table\u0026nbsp;20)\u003c/b\u003e.\u003c/p\u003e \u003cp\u003eWe quantified the total microbiome abundance of each enzyme across the four menstrual phases (see \u003cb\u003eMethods\u003c/b\u003e) using HUMAnN \u003cb\u003e(Supplementary Fig.\u0026nbsp;9)\u003c/b\u003e. We found that throughout all phases, beta-galactosidase was the most prevalent, while aryl sulfate sulfotransferase exhibited the lowest abundance (\u003cb\u003eSupplementary Fig.\u0026nbsp;9\u003c/b\u003e). We then examined the gene abundance differences across pairs of menstrual phases using paired Wilcoxon signed-rank tests (\u003cb\u003eSupplementary Fig.\u0026nbsp;10)\u003c/b\u003e, and observed that after multiple testing correction there were significantly lower levels of beta-galactosidase, beta-glucosidase, arylsulfatase and beta-glucuronidase in follicular versus early luteal phase (q-values 0.01, 0.05, 0.13 and 0.22, respectively). Furthermore, beta-galactosidase abundance was also lower in the ovulatory phase compared to follicular phase (q\u0026thinsp;=\u0026thinsp;0.22) \u003cb\u003e(Supplementary Table\u0026nbsp;21).\u003c/b\u003e\u003c/p\u003e \u003cp\u003eNext, we evaluated the association between enzyme abundances and menstrual phases, and between enzyme abundances and reproductive hormones (17β-estradiol and progesterone) using linear mixed-effects models. However, none of the enzymes showed significant change across phases (\u003cb\u003eSupplementary Table\u0026nbsp;22\u003c/b\u003e) or associations with hormones (\u003cb\u003eSupplementary Table\u0026nbsp;23\u003c/b\u003e).\u003c/p\u003e \u003cp\u003eTo evaluate the overall contribution of enzyme abundance changes to reproductive hormone levels, we compared the performance of a basic model that included progesterone (or 17β-estradiol ) as response variable and host covariates as explanatory variables (menstrual phase, BMI, age, Bristol stool scale, time of sample collection and technical covariates) with a complete model that incorporated these variables as well as the six enzyme abundances. For progesterone, the complete model did not improve the fit compared to the base model (p\u0026thinsp;=\u0026thinsp;0.56). Moreover, model selection criteria indicated no improvement, with higher AIC (1492 vs 1476) and BIC (1562 vs 1536) values observed for the complete model. Similarly, for 17β-estradiol, the inclusion of the gene enzymes did not improve the model performance (p\u0026thinsp;=\u0026thinsp;0.55). Again, the complete model showed higher AIC (1116 vs 1103) and BIC (1199 vs 1154) values compared to the base model.\u003c/p\u003e \u003c/div\u003e"},{"header":"Discussion","content":"\u003cp\u003eThere is currently a significant knowledge gap regarding gut microbiome composition and dynamics across a natural menstrual cycle in women of reproductive age without gynecological or metabolic conditions. The present study addressed this gap while additionally examining dietary parameters and reproductive hormone fluctuations during the menstrual cycle.\u003c/p\u003e \u003cp\u003e To assess the contribution of diet to microbiome variability, we first evaluated the consistency in dietary intake estimates obtained by two sources of dietary information: food frequency questionnaires (FFQs) and daily food diaries. Comparison of nutrients estimated by these two approaches revealed significant discrepancies consistent with known bias in subjective surveys \u003csup\u003e\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e\u003c/sup\u003e. In particular, daily food diaries - which likely provide a more accurate reflection of actual food consumption compared to a retrospective assessment - provided higher estimates of total energy and macronutrients, but lower estimates of cholesterol and fiber compared to the FFQ. These differences may arise from several aspects, including limited dietary self-awareness, the simplified FFQ used in this study, or potential errors and lack of country-specific data in the MyFitnessPal database as previously reported \u003csup\u003e\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e,\u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e\u003c/sup\u003e. Nevertheless, despite absolute differences in nutrient estimates, nutrients derived from the two methods showed significant positive correlation. This indicates a generally consistent ranking of individuals across assessment approaches, supporting the complementary use of FFQs to capture habitual food consumption and daily food diaries to reflect nutrient intake proximal to microbiome sampling.\u003c/p\u003e \u003cp\u003e We next examined associations between dietary components and gut microbiome composition and function, using dietary parameters obtained from both types of information. Several of the observed associations replicate previously reported findings, reinforcing the validity of our analyses. For example, we confirmed the positive association between fiber and fruit intake with abundance of \u003cem\u003eRoseburia hominis\u003c/em\u003e, a known short-chain fatty acid (SCFA) producer, capable of breaking down dietary fibers \u003csup\u003e\u003cspan additionalcitationids=\"CR45\" citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e\u003c/sup\u003e. We also confirmed the well-established positive association between coffee consumption and \u003cem\u003eLawsonibacter asaccharolyticus\u003c/em\u003e\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u003c/sup\u003e. At the same time, though, we observed a novel association of \u003cem\u003eL\u003c/em\u003e. \u003cem\u003easaccharolyticus\u003c/em\u003e with yogurt intake that was independent of coffee consumption, suggesting that this species may respond more broadly to fermented food consumption. Furthermore, we identified additional species belonging to the \u003cem\u003eAlistipes\u003c/em\u003e genus (e.g. \u003cem\u003eA. ihumii and A. communis\u003c/em\u003e) to be associated with coffee consumption, a genus that has previously been associated with coffee and caffeine intake \u003csup\u003e\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e\u003c/sup\u003e. Among the novel findings, \u003cem\u003eAlistipes shahii\u003c/em\u003e and \u003cem\u003eHominenteromicrobium mulieris\u003c/em\u003e were associated with dietary cholesterol intake. Considering that the genus \u003cem\u003eAlistipes\u003c/em\u003e is receiving growing attention in dyslipidemic conditions, and that \u003cem\u003eH. mulieris\u003c/em\u003e has been recently associated with intestinal mucosal health \u003csup\u003e\u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u003c/sup\u003e, these associations merit further investigation to clarify whether high dietary cholesterol selectively promotes microbial taxa with potential beneficial or compensatory functions. Another novel relationship involved \u003cem\u003eParabacteroides merdae\u003c/em\u003e, a species known to metabolize branched-chain amino acids (BCAAs) \u003csup\u003e\u003cspan citationid=\"CR48\" class=\"CitationRef\"\u003e48\u003c/span\u003e\u003c/sup\u003e, which was associated with cold cuts consumption, foods typically rich in BCAA. Diet also influenced microbiome function, which we capture using microbial pathways abundances derived from our shot-gut sequencing approach. Among worth noting associations, is the inverse relationship between cholesterol intake and pathways involved in bacterial lipid biosynthesis. Consistent with this observation, prior animal studies have reported correlations between fasting triglyceride levels and bacterial lipid-related pathways, including CMP 3 deoxy D manno-octulosonate biosynthesis \u003csup\u003e\u003cspan citationid=\"CR49\" class=\"CitationRef\"\u003e49\u003c/span\u003e\u003c/sup\u003e. Collectively, these findings reinforce the notion that both individual foods and general dietary patterns influence gut bacterial composition and function.\u003c/p\u003e \u003cp\u003eSex is another host factor consistently associated with gut microbiome interindividual variability. Several studies have shown that young women show greater microbial diversity than age-matched men; however this difference diminishes with age and is no longer detectable after menopause \u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e,\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e,\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e\u003c/sup\u003e. This pattern has been largely attributed to the decline in estrogen production during menopause and its potential interaction with gut bacteria involved in estrogen metabolization, collectively referred to as the estrobolome \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. Microbial taxa shown to decrease after menopause include species from \u003cem\u003ePrevotella, Bacteroides, Parabacteroides\u003c/em\u003e., as well as \u003cem\u003eA. muciniphila\u003c/em\u003e, and their abundance strongly correlated with depleted gut microbial β-glucuronidases enzymes \u003csup\u003e\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e\u003c/sup\u003e. Despite this evidence, the contribution of other reproductive hormones, many of which undergo substantial changes during menopause, remain largely unexplored in both pre- and post-menopause women.\u003c/p\u003e \u003cp\u003eIn this context, we assessed the impact of menstrual cycle phases and five reproductive hormones on the gut microbiome. Intriguingly, microbiome changes were visible and significant already on overall microbiome metrics such as nd richness and diversity. Richness was modestly higher during the ovulatory phase compared to the follicular phase. Differences in beta diversity were instead more pronounced, with the greatest interindividual variation observed during the late luteal phase compared to the follicular phase. This suggests that proximity to menstruation may be associated with increased microbial variability between individuals.\u003c/p\u003e \u003cp\u003eAt higher resolution, we identified microbial species and microbial pathways whose abundances were associated with a menstrual phase or with circulating reproductive hormones or both. Some of these associations align with previous findings on patients, thus bringing additional support to the detected relationships. For example, progesterone levels were positively associated with \u003cem\u003eCollinsella aerofaciens\u003c/em\u003e. Abundance of this species was previously shown to increase in endometriosis patients following treatment with Dienogest, a progestin medication \u003csup\u003e\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e\u003c/sup\u003e. Intriguingly, most of our associations were detected with FSH, LH and progesterone rather than with 17β-estradiol. The association with FSH and LH is likely indirect, potentially via the gut-brain axis, signaling to hypothalamus and pituitary glands \u003csup\u003e\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e\u003c/sup\u003e. The gut-brain axis may also explain the few associations between microbial species and pathways with prolactin levels, prolactin being secreted under inhibitory control by hypothalamic dopaminergic neurons. Finally, a subset of species associated with hormones were also associated with dietary factors, but we were unable to find evidence for mediation of either diet on gut-hormones associations or of hormones on diet-gut associations. Despite only a few associations being observed with 17β-estradiol, we noted that several included bacterial species potentially involved in the estrobolome activity, particularly members of the \u003cem\u003eBacteroides\u003c/em\u003e genus. While the underlying mechanisms of estrobolome have been widely described for estrogens, they could potentially be extended to progesterone \u003csup\u003e\u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e53\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eWe thus quantified the abundance of genes encoding estrobolome enzymes and investigated their association with circulating 17β-estradiol and progesterone levels. Intriguingly, we see menstrual cycle associated changes in abundance of the most relevant enzymes: beta-galactosidase, beta-glucosidase, and higher arylsulfatase type1 and beta-glucuronidase. None of these were however associated with circulating hormones, a negative result that may be attributed to the fact that we are unable to distinguish systemic and de-conjugated fractions, which may more directly reflect microbial activity, or to the fact that we measured hormones in plasma rather than in feces. Indeed, while mechanistic studies suggest the steroid metabolizing capacity of the gut microbiome, human studies have so far yielded inconsistent results regarding association between estrobolome enzymes and plasma circulating estrogen levels \u003csup\u003e\u003cspan citationid=\"CR54\" class=\"CitationRef\"\u003e54\u003c/span\u003e,\u003cspan citationid=\"CR55\" class=\"CitationRef\"\u003e55\u003c/span\u003e\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eIn parallel, our data highlight the importance of SCFA-producing bacteria in the context of hormonal regulation. In fact, a total of 21 out of the 30 species associated with hormone and diet are also SCFA producers, including the well-characterized \u003cem\u003eAkkermansia muciniphila\u003c/em\u003e and \u003cem\u003eFaecalibacterium prausnitzii\u003c/em\u003e. While beta-glucuronidase producing bacteria facilitate estrogen recirculation via enterohepatic recycling, we hypothesize that SCFA producing taxa may exert complementary effects by first modulating host metabolism and, in turn, impact endocrine responses and hormonal homeostasis. Direct measurement of SCFA also in future expansions of this study are needed to clarify this hypothesis.\u003c/p\u003e \u003cp\u003eTo our knowledge, this study provides the first evidence of microbiome variations during menstrual cycle in healthy women with concurrent hormone measurements, and the first to interrogate these changes in relation to reproductive hormones beyond estrogen and progesterone. We also acknowledge its limitations. First, the sample size is still relatively small, and thus we may be underpowered to pinpoint smaller effects of hormones on microbiome. Furthermore, our study includes only healthy women, in whom physiological hormonal changes occur, so microbiome\u0026ndash;hormone interactions are likely too subtle to fully be captured. Second, we measured only total circulating hormones, without performing specific assays to differentiate between conjugated forms, directly synthesized unconjugated forms, or unconjugated forms resulting from deconjugation. Third, volunteers exhibited relatively homogeneous dietary patterns, limiting our power to detect potential hormonal mediation of diet\u0026ndash;microbiome interactions. Finally, we acknowledge that other unmeasured factors, such as sleep quality and stress, could also influence microbiome dynamics during the menstrual cycle.\u003c/p\u003e \u003cp\u003eIn conclusion, our findings indicate that physiological fluctuation in endogenous reproductive hormones across a natural menstrual cycle are associated with measurable changes in gut microbiome composition and function, and their effects are distinct from dietary factors. These results underscore the importance of incorporating sex-specific data and hormonal status into microbiome research. Our study provides novel insights into host-microbiome interactions that are specific to female physiology, highlighting a strong relationship with gonadotropics hormones. Future research will be necessary to build upon these observations and clarify their clinical relevance, especially for reproductive health. In particular, larger longitudinal cohorts integrating multi-omics approaches will be critical to elucidate how gonadotropins together with sex steroids interact with microbial metabolism and host signaling pathways. Such efforts will help determine whether hormonally-associated microbiome changes translate into measurable physiological consequences and whether they can be leveraged to inform personalized, menstrual cycle-aware interventions aimed at improving women\u0026rsquo;s health.\u003c/p\u003e \u003cdiv id=\"Sec26\" class=\"Section2\"\u003e \u003ch2\u003eSupplementary Notes File (PDF)\u003c/h2\u003e \u003cp\u003eIn the Supplementary Notes we describe how dietary data were collected, processed, and analyzed in the study. The file provides details on the transformation of FFQ data into nutrient intakes and diet scores, and reports differences in nutrient intake, diet scores, and gut microbiome composition across centers and hormonal phases. Supplementary Notes figures and tables are included along with their legend.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec27\" class=\"Section2\"\u003e \u003ch2\u003eSupplementary Figures (PDF)\u003c/h2\u003e \u003cp\u003eIn the Supplementary Figures file we provide Supplementary Fig.\u0026nbsp;1 to 10. Supplementary Figures legends are included in the file.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec28\" class=\"Section2\"\u003e \u003ch2\u003eSupplementary Tables (xlsx)\u003c/h2\u003e \u003cp\u003eThis file contains 23 Supplementary Tables and one tab containing a summary.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec29\" class=\"Section2\"\u003e \u003ch2\u003eData availability statement\u003c/h2\u003e \u003cp\u003eAll our code used to analyse the data and instructions are available at: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://github.com/Sanna-s-LAB/Women4Health/tree/main/Code_for_LoFaro_et_al\u003c/span\u003e\u003cspan address=\"https://github.com/Sanna-s-LAB/Women4Health/tree/main/Code_for_LoFaro_et_al\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/p\u003e \u003cp\u003eWe provide relative abundance of species and pathways in Zenodo along with information on visit number and menstrual phase at the following DOI: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.5281/zenodo.19953398\u003c/span\u003e\u003cspan address=\"10.5281/zenodo.19953398\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. The data will be made public after article acceptance. The temporary link for reviewers is the following: \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://zenodo.org/records/19953398?preview=1\u0026amp;token=eyJhbGciOiJIUzUxMiJ9.eyJpZCI6IjRmYjljODEzLTRiZTItNDVhMi05MjY4LWY1M2UyMmVlNGY1NSIsImRhdGEiOnt9LCJyYW5kb20iOiIwZTQwN2QzZTNlZjYyMzNkMTMzNDg4NmRkZmYyZGJhZCJ9.8wKj4f4cK-eZS1DEnbaiLOqFL-wt4L3LTJjpTz5j3vBvCT6t4Jt4iGfGIq6POgtmzBz8txRrnVyvCv5FzgsHpg\u003c/span\u003e\u003cspan address=\"https://zenodo.org/records/19953398?preview=1\u0026amp;token=eyJhbGciOiJIUzUxMiJ9.eyJpZCI6IjRmYjljODEzLTRiZTItNDVhMi05MjY4LWY1M2UyMmVlNGY1NSIsImRhdGEiOnt9LCJyYW5kb20iOiIwZTQwN2QzZTNlZjYyMzNkMTMzNDg4NmRkZmYyZGJhZCJ9.8wKj4f4cK-eZS1DEnbaiLOqFL-wt4L3LTJjpTz5j3vBvCT6t4Jt4iGfGIq6POgtmzBz8txRrnVyvCv5FzgsHpg\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/p\u003e \u003cp\u003eOther participant-level personally identifiable data cannot be shared in respect to participants' informed consent and are protected under the Italian Personal Data Protection Code and European Regulation 2016/679 of the European Parliament and of the Council (GDPR) that prohibit distribution even in pseudo-anonymised form.\u003c/p\u003e \u003cp\u003eAll data necessary to support the conclusions drawn in this study are available in the Supplementary Tables.\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003eThe study protocol - including informed consent forms - was approved by the Friuli-Venezia Giulia Ethical Committee on 09/2021 with prot. N. 0034184/P/GEN/ARCS and modifications amended on 09/2023 with prot. N. 0033477/P/GEN/ARCS; by the Sardinian Ethical Committee on 05/23 with prot. N. 1.1 and modifications amended on 07/2024 with prot. N.1.2; and by the Area Vasta Emilia Romagna Ethical Committee on 12/2024 with prot. N 1875/2024.\u003c/p\u003e\u003cp\u003e \u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eThis study was co-funded by the European Union (ERC Stg 2022 to S.S., acronym SEMICYCLE, GA n.101075624), by the Next Generation EU, in the context of the National Recovery and Resilience Plan, Investment PE8 \u0026ndash; Project Age-It: \u0026ldquo;Ageing Well in an Ageing Society \u0026rdquo; [DM 1557 11.10.2022 to S.S.], by NutrAGE grant (CNR Project FOE-2021 DBA.AD005.225 to S.S.), by InvAT grant (CNR Project FOE-2022 DSB.AD006.371.001 to S.S.) and by the Italian Ministry of Health through the contribution given to the Institute for Maternal and Child Health IRCCS Burlo Garofolo, Trieste - Italy (SD 02/21 to G.G.). Views and opinions expressed are, however, those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Council. Neither the European Union nor the granting authority can be held responsible for them. In addition, S.S. also received a PRIN2022 grant from the Next Generation EU funds (DSB.PN004.021 2022PMZKEC_LS2_PRIN2022 SANNA, CUP B53D23008300006).\u003c/p\u003e \u003c/p\u003e \u003cp\u003e \u003cstrong\u003eUse of AI statement\u003c/strong\u003e \u003cp\u003eThe authors declare that they have used generative artificial intelligence, ChatGPT, Quillbot (2026) and Copilot online platforms only to check spelling, grammar and expressions, adapting it to an academic scientific style.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eFunding\u003c/h2\u003e \u003cp\u003eacquisition: G.G., S.S.\u003c/p\u003e \u003cp\u003eAll authors read and revised the manuscript and approved its final version. Furthermore, all authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.\u003c/p\u003e\u003ch2\u003eAuthors\u0026rsquo; contributions\u003c/h2\u003e \u003cp\u003eStatistical and Bioinformatic analyses: V.LoF., P.F., A.C., D.V.Z., S.S.\u003c/p\u003e\u003ch2\u003eAcknowledgements\u003c/h2\u003e \u003cp\u003e We warmly thank all the volunteers who participated in the study for their time and commitment. We also thank all collaborators involved in the recruitment phase, in particular: Dr. Giovanni di Lorenzo and the other physicians of the S.C. Ostetricia e Ginecologia of the IRCCS Burlo-Garofolo, Dr. Daniela Mazz\u0026agrave; of the S.C. Genetics of the IRCCS Burlo-Garofolo, nurses Maria Paola Fercia and Maria Paola Orani, technician Siro Agus and Dr. Elena Poma at the Azienda Ospedaliero Universitaria di Cagliari, and Alice Cordani, Francesca Bianco and Maria Chiara Matteucci at the University of Bologna. More, we thank the Directors of IRCCS Burlo-Garofolo and of IRGB-CNR for logistic support, the Agenzia Regionale Sardegna Ricerche for providing access to laboratories, Dr Stefania Olla from IRGB-CNR for her critical feedback on the enzymes analysis, technologists Marco Masala and Michele Marongiu from the IRGB-CNR for IT support. Finally, we gratefully acknowledge our colleague, Maria Laura Ferrando, who passed away during the preparation of this work and contributed to it with enthusiasm until the end.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eCarlino N et al (2024) Unexplored microbial diversity from 2,500 food metagenomes and links with the human microbiome. Cell 187:5775\u0026ndash;5795e15\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAsnicar F et al (2021) Microbiome connections with host metabolism and habitual diet from 1,098 deeply phenotyped individuals. Nat Med 27:321\u0026ndash;332\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAsnicar F et al (2026) Gut micro-organisms associated with health, nutrition and dietary interventions. Nature 650:450\u0026ndash;458\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCampaniello D et al (2022) How Diet and Physical Activity Modulate Gut Microbiota: Evidence, and Perspectives. Nutrients 14:2456\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFackelmann G et al (2025) Gut microbiome signatures of vegan, vegetarian and omnivore diets and associated health outcomes across 21,561 individuals. Nat Microbiol 10:41\u0026ndash;52\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eManghi P et al (2024) Coffee consumption is associated with intestinal Lawsonibacter asaccharolyticus abundance and prevalence across multiple cohorts. Nat Microbiol 9:3120\u0026ndash;3134\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFalony G et al (2016) Population-level analysis of gut microbiome variation. Science 352:560\u0026ndash;564\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGacesa R et al (2022) Environmental factors shaping the gut microbiome in a Dutch population. Nature 604:732\u0026ndash;739\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFerrando ML, Busonero F, Crobu F, Sanna S (2026) Aging in women \u0026ndash; The microbiome perspective. Ageing Res Rev 113:102950\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang X et al (2021) Sex- and age-related trajectories of the adult human gut microbiota shared across populations of different ethnicities. Nat Aging 1:87\u0026ndash;100\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ede la Cuesta-Zuluaga J et al (2019) Age- and Sex-Dependent Patterns of Gut Microbial Diversity in Human Adults. mSystems 4:e00261\u0026ndash;e00219\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDabek M, McCrae SI, Stevens VJ, Duncan SH, Louis P (2008) Distribution of beta-glucosidase and beta-glucuronidase activity and of beta-glucuronidase gene gus in human colonic bacteria. FEMS Microbiol Ecol 66:487\u0026ndash;495\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eErvin SM et al (2020) Structural Insights into Endobiotic Reactivation by Human Gut Microbiome-Encoded Sulfatases. Biochemistry 59:3939\u0026ndash;3950\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eErvin SM et al (2019) Gut microbial β-glucuronidases reactivate estrogens as components of the estrobolome that reactivate estrogens. J Biol Chem 294:18586\u0026ndash;18599\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLarnder AH, Manges AR, Murphy RA (2025) The estrobolome: Estrogen-metabolizing pathways of the gut microbiome and their relation to breast cancer. Int J Cancer 157:599\u0026ndash;613\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePeters BA et al (2022) Menopause Is Associated with an Altered Gut Microbiome and Estrobolome, with Implications for Adverse Cardiometabolic Risk in the Hispanic Community Health Study/Study of Latinos. \u003cem\u003emSystems\u003c/em\u003e 7, e0027322\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLey D, Saha S (2025) Menopause and gastrointestinal health and disease. Nat Rev Gastroenterol Hepatol 22:556\u0026ndash;569\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCross T-WL et al (2024) Gut microbiome responds to alteration in female sex hormone status and exacerbates metabolic dysfunction. Gut Microbes 16:2295429\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBermingham KM et al (2022) Menopause is associated with postprandial metabolism, metabolic health and lifestyle: The ZOE PREDICT study. eBioMedicine 85\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePark MG, Cho S, Oh MM (2023) Menopausal Changes in the Microbiome\u0026mdash;A Review Focused on the Genitourinary Microbiome. Diagnostics 13\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrito J et al (2025) Hormonal birth control is associated with altered gut microbiota β-diversity in physically active females across the menstrual cycle: a pilot trial. J Appl Physiol 138:739\u0026ndash;745\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi VW et al (2024) Mass spectrometric profiling of estrogen and estrogen metabolites in human stool and plasma partially elucidates the role of the gut microbiome in estrogen recycling. 08.07.606543 Preprint at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1101/2024.08.07.606543\u003c/span\u003e\u003cspan address=\"10.1101/2024.08.07.606543\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2024)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVandeputte D et al (2021) Temporal variability in quantitative human gut microbiome profiles and implications for clinical research. Nat Commun 12:6740\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBusonero F et al (2024) The Women4Health cohort: a unique cohort to study women-specific mechanisms of cardio-metabolic regulation. Eur Heart J Open 4:oeae012\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSanna S et al (2026) Longitudinal plasma proteomics along a menstrual cycle highlights the regulatory effect of sex hormones on inflammatory and cardiometabolic circuits. Preprint at. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.21203/rs.3.rs-9148671/v1\u003c/span\u003e\u003cspan address=\"10.21203/rs.3.rs-9148671/v1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLangmead B, Salzberg SL (2012) Fast gapped-read alignment with Bowtie 2. Nat Methods 9:357\u0026ndash;359\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlanco-M\u0026iacute;guez A et al (2023) Extending and improving metagenomic taxonomic profiling with uncharacterized species using MetaPhlAn 4. Nat Biotechnol 41:1633\u0026ndash;1644\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBeghini F et al (2021) Integrating taxonomic, functional, and strain-level profiling of diverse microbial communities with bioBakery 3. \u003cem\u003eeLife\u003c/em\u003e 10, e65088\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHu S et al Gut microbial beta-glucuronidase: a vital regulator in female estrogen metabolism. Gut Microbes 15, 2236749\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGodse R, Fernandes JM, Kulkarni R (2025) Characterization of β-glucosidase activity of a Lactiplantibacillus plantarum 6-phospho-β-glucosidase. Appl Microbiol Biotechnol 109:86\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang W et al (2025) Roseburia hominis improves host metabolism in diet-induced obesity. Gut Microbes 17:2467193\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Y, Walker RW, Kaplan RC, Qi Q (2025) Added sugars, gut microbiota, and host health. Gut Microbes 17:2592431\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang Q et al (2020) Role of Dietary Nutrients in the Modulation of Gut Microbiota: A Narrative Review. Nutrients 12:381\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJumas-Bilak E et al (2007) Acidaminococcus intestini sp. nov., isolated from human clinical samples. Int J Syst Evol Microbiol 57:2314\u0026ndash;2319\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDai A et al (2023) The Association between Caffeine Intake and the Colonic Mucosa-Associated Gut Microbiota in Humans-A Preliminary Investigation. Nutrients 15:1747\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLe Roy CI et al (2022) Yoghurt consumption is associated with changes in the composition of the human gut microbiome and metabolome. BMC Microbiol 22:39\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYun C et al (2024) The microbial metabolite agmatine acts as an FXR agonist to promote polycystic ovary syndrome in female mice. Nat Metab 6:947\u0026ndash;962\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTorres PJ et al (2018) Gut Microbial Diversity in Women With Polycystic Ovary Syndrome Correlates With Hyperandrogenism. J Clin Endocrinol Metab 103:1502\u0026ndash;1511\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChu W et al (2020) Metagenomic analysis identified microbiome alterations and pathological association between intestinal microbiota and polycystic ovary syndrome. Fertil Steril 113:1286\u0026ndash;1298e4\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSakamuri A et al (2023) Sex hormones, sex chromosomes, and microbiota: Identification of Akkermansia muciniphila as an estrogen-responsive microbiota. Microbiota Host 1:e230010\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eShim J-S, Oh K, Kim HC (2014) Dietary assessment methods in epidemiologic studies. Epidemiol Health 36:e2014009\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTeixeira V, Voci SM, Mendes-Netto RS (2018) \u0026amp; da Silva, D. G. The relative validity of a food record using the smartphone application MyFitnessPal. Nutr Diet 75:219\u0026ndash;225\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTosi M et al (2021) Accuracy of applications to monitor food intake: Evaluation by comparison with 3-d food diary. Nutrition 84:111018\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSong L et al (2022) Roseburia hominis Alleviates Neuroinflammation via Short-Chain Fatty Acids through Histone Deacetylase Inhibition. Mol Nutr Food Res 66:e2200164\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTamanai-Shacoori Z et al (2017) Roseburia spp.: a marker of health? Future Microbiol 12:157\u0026ndash;170\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSegev T et al (2026) Diet-microbiome associations in 10,068 individuals from the Human Phenotype Project to guide personalized nutrition. Nat Med. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1038/s41591-026-04312-x\u003c/span\u003e\u003cspan address=\"10.1038/s41591-026-04312-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eCu\u0026iacute;v P\u0026Oacute; et al (2025) A human microbiome-derived therapeutic for ulcerative colitis promotes mucosal healing and immune homeostasis. 04.18.25325957 Preprint at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1101/2025.04.18.25325957\u003c/span\u003e\u003cspan address=\"10.1101/2025.04.18.25325957\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2025)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang H et al (2025) Gut microbiota metabolism of branched-chain amino acids and their metabolites can improve the physiological function of aging mice. Aging Cell 24:e14434\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHorne RG et al (2020) High Fat-High Fructose Diet-Induced Changes in the Gut Microbiota Associated with Dyslipidemia in Syrian Hamsters. Nutrients 12:3557\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSantos-Marcos JA et al (2018) Influence of gender and menopausal status on gut microbiota. Maturitas 116:43\u0026ndash;53\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePronina V et al (2025) Gut Microbial Composition on Dienogest Therapy in Patients with Endometriosis. Microbiol Res 16\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eOrganski AC, Jorgensen JS, Cross T-WL (2021) Involving the life inside: The complex interplay between reproductive axis hormones and gut microbiota. Curr Opin Endocr Metabolic Res 20:100284\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChapadgaonkar SS, Bajpai SS, Godbole MS (2023) Gut microbiome influences incidence and outcomes of breast cancer by regulating levels and activity of steroid hormones in women. Cancer Rep (Hoboken) 6:e1847\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLarnder AH et al (2026) Gut microbiome and metabolome reveal hormone-related and functional alterations in ER-positive breast cancer: a case\u0026ndash;control study. 02.06.26345778 Preprint at \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.64898/2026.02.06.26345778\u003c/span\u003e\u003cspan address=\"10.64898/2026.02.06.26345778\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e (2026)\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNannini G, Cei F, Amedei A (2025) Unraveling the Contribution of Estrobolome Alterations to Endometriosis Pathogenesis. Curr Issues Mol Biol 47\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-9888793/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9888793/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eSex differences in the gut microbiome are well established, yet the contribution of reproductive hormone variations within women remains insufficiently characterized. Here, we investigated how hormonal fluctuations during a natural menstrual cycle shape gut microbiome in healthy women and whether these effects are distinct from dietary influences. Leveraging longitudinal shot-gun gut microbiome data from 160 Italian participants in the Women4Health cohort, combined with detailed dietary information, we examined relationships with five reproductive hormones hormones: 17β-estradiol, progesterone, follicle-stimulating hormone (FSH), luteinizing hormone (LH) and prolactin. Microbial alpha diversity was significantly lower at the follicular phase compared to ovulatory phase (p\u0026thinsp;=\u0026thinsp;0.03), while beta diversity significantly increased from follicular to the late luteal phase (p\u0026thinsp;=\u0026thinsp;8.7x10\u003csup\u003e\u0026minus;\u0026thinsp;6\u003c/sup\u003e). Twenty-eight species and twenty-two microbial pathways were significantly associated with reproductive hormones, with the majority of associations observed for gonadotropins (FSH and LH) rather than gonadal steroids (17β-estradiol and progesterone). While we confirm strong relationships between diet and microbiome, mediation analysis showed no evidence for a mediating effect of hormones on those. Together, these findings reveal distinct hormonal and dietary influences on gut microbiome across the menstrual cycle in healthy women.\u003c/p\u003e","manuscriptTitle":"Gut Microbiome Variation Across Menstrual Phases: Distinct Roles of Diet and Physiological Hormonal Fluctuations in Healthy Women","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-06-16 15:20:48","doi":"10.21203/rs.3.rs-9888793/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-communications","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"NCOMMS","sideBox":"Learn more about [Nature Communications](http://www.nature.com/ncomms/)","snPcode":"","submissionUrl":"https://mts-ncomms.nature.com/","title":"Nature Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Communications","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"eaed14dc-0211-43a5-bc66-793a952d956b","owner":[],"postedDate":"June 16th, 2026","published":true,"recentEditorialEvents":[{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-07-17T15:01:25+00:00","index":3,"fulltext":"This content is not available."}],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":69513153,"name":"Biological sciences/Computational biology and bioinformatics/Data processing"},{"id":69513154,"name":"Health sciences/Medical research/Epidemiology"},{"id":69513155,"name":"Biological sciences/Microbiology/Microbial genetics/Bacterial genes"},{"id":69513156,"name":"Health sciences/Risk factors"}],"tags":[],"updatedAt":"2026-06-16T15:20:48+00:00","versionOfRecord":[],"versionCreatedAt":"2026-06-16 15:20:48","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9888793","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9888793","identity":"rs-9888793","version":["v1"]},"buildId":"ehx78VzkSd0WSzXnipQa-","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}