Polyphenol-Related Gut Metabotype Signatures Linked to Quality of Life in Postmenopausal Women: A Randomized, Placebo-Controlled Crossover Trial

High-performance liquid chromatography (HPLC)-grade acetonitrile, dimethyl sulfoxide (DMSO), formic acid, and methanol were obtained from JT Baker (Deventer, The Netherlands). trans-Resveratrol (resveratrol, RSV), ellagic acid (EA), daidzein, genistein, formononetin, biochanin A, equol, and hesperetin were purchased from Sigma-Aldrich (St. Louis, MO, USA). Equol 7-O-glucuronide and O-demethylangolensin (ODMA) were purchased from LGC Standards (Barcelona, Spain). RSV 4′-O-sulfate, RSV 3-O-glucuronide, dihydroresveratrol 3-O-glucuronide, RSV 3-O-sulfate, Urolithin A (Uro-A), B (Uro-B), Isourolithin A (IsoUro-A), Uro-A 3-O-glucuronide, Uro-A 8-O-glucuronide, Uro-B glucuronide, Uro-A sulfate, Uro-B sulfate, IsoUro-A 3-O-glucuronide, IsoUro-A 9-O-glucuronide, lunularin (LUNU), 4-hydroxydibenzyl (4HDB), LUNU glucuronide, LUNU sulfate, 4HDB glucuronide, and 4HDB sulfate were obtained as described elsewhere [25,34,35]. All the reagents and metabolites were 97% pure or higher. Ultrapure Millipore water (Darmstadt, Germany) was used throughout the study.

The polyphenol-rich supplement (PPs) consisted of a mixture of RSV and commercially available pomegranate and red clover extracts. The components were mixed and encapsulated in hard gelatin capsules by Laboratorios Admira S.L. (Alcantarilla, Murcia, Spain). The capsules were manufactured, tested, and checked in accordance with the European Union's Good Manufacturing Practices requirements. Each capsule (700 mg) contained a mixture of 50 mg trans-RSV (98% purity) from Polygonum cuspidatum, 320 mg punicalagin-rich pomegranate extract (20% punicalagin), 80 mg EA-rich pomegranate extract (40% EA), and 250 mg red clover extract (20% isoflavones). The PPs mixture was analyzed by High-Performance Liquid Chromatography coupled to Electrospray Ionization-Ion Trap Tandem Mass Spectrometry (HPLC-ESI-IT-MS/MS), as previously described [36], and its detailed composition is shown in Table 1. The placebo capsules (Pla) contained microcrystalline cellulose (700 mg) and were indistinguishable from those containing the extract mixture.

The study protocol was approved by the Bioethics Committee of the Spanish National Research Council (CSIC) (Madrid, Spain) (reference 249/2023), IMDEA-Food (Madrid) (reference PI-065), and the Virgen de La Arrixaca University Hospital (Murcia, Spain) (reference 2023-4-9-HCUVA) within the PolyPAUSE Project (PID2022-136419OB-I00; MICIU, Spain). The trial adhered to the principles of good clinical practice outlined in the Declaration of Helsinki of 1975 and its subsequent amendments, and it was registered at clinicaltrials.gov (NCT07182370). Participants were enrolled from September 2023 to January 2025.

The trial consisted of two studies: Study 1 involved women of reproductive age (premenopausal; Pre-M) with a 3-day follow-up for metabotype stratification, and Study 2 included Post-M women with an initial 3-day follow-up for metabotype stratification and an 8-week follow-up to assess the impact of PPs on QoL.

Study 1. Healthy Pre-M women (aged 30−48 years) were recruited at the IMDEA Food Institute (Madrid) and CEBAS-CSIC (Murcia, Spain). Exclusion criteria were as follows: intake of antibiotics within the month before the study; use of chronic medication or any other medication, even sporadically, within one month before the start of the study; BMI < 18 kg/m²; swallowing difficulties; pregnancy or lactation; smoking; diagnosed chronic illness; previous gastrointestinal surgery; habitual alcohol consumption > 15 g/day; vegetarian diet; allergy, intolerance, or regular consumption of red clover extract, resveratrol (grape, wine), soy, or pomegranate; or current participation in a weight-loss regimen. The study was fully explained to the volunteers, who provided written informed consent before participation. Physical activity levels were assessed using the validated International Physical Activity Questionnaire (IPAQ) [37]. Adherence to the Mediterranean diet was recorded using the validated PREDIMED questionnaire [38].

Pre-M participants consumed three capsules per day in the evening, containing the polyphenol-rich extract mixture (PPs), for three consecutive days. Participants were instructed to take the capsules in the evening to standardize for potential circadian effects and, in particular, to maximize the likelihood of detecting polyphenol-derived metabolites in the first morning urine sample after fasting. The daily intake of PPs (3 capsules) provided 133.2 ± 10.1 mg of RSV, 312.0 ± 30.9 mg of ellagitannins (including punicalin, α- and β-punicalagins) and free ellagic acid, as well as 166.3 ± 27.3 mg of isoflavones, including daidzein, genistein, formononetin, and biochanin A. At the end of the study, the participants provided the questionnaires and a urine sample. The aim was to stratify this group of women according to their polyphenol metabotype and corresponding clusters (MCs) [22].

Study 2. Healthy postmenopausal women (Post-M) were recruited at IMDEA Food, CEBAS-CSIC, and the Virgen de La Arrixaca University Hospital (Murcia). Inclusion criteria were the following: women aged 45–59 years, with menopause (defined as ≥1 year without menstruation), presenting at least one climacteric symptom (hot flashes, sweating episodes, low mood, irritability, altered libido, insomnia, and/or joint or muscle pain), and with a BMI > 18 kg/m². Exclusion criteria were as follows: chronic-degenerative or psychiatric conditions; medical contraindications to nutritional supplements; chronic medication use (e.g., cholesterol, blood pressure, hypertension, anxiety, insomnia, or others) or any other medication, even sporadically, within one month before the start of the study; antibiotic use within the month before enrollment; hormone replacement therapy; history of major gastrointestinal surgery; swallowing difficulties; BMI < 18 kg/m²; current participation in a weight-loss regimen; allergy or intolerance to red clover extract, resveratrol (grape, wine), soy, or pomegranate; smoking; habitual alcohol consumption > 15 g/day; vegetarian diet; or regular intake of supplements containing resveratrol, pomegranate, or isoflavones. Participants received a comprehensive explanation of the study and provided written informed consent before enrollment. Physical activity level and adherence to the Mediterranean diet were also recorded as described above. The primary outcome of the trial was the distribution of metabotypes and their combinations (i.e., metabotype clusters, MCs) in Post-Mwomen. A sample size of 90 participants was considered based on prior studies and the logistical difficulty of recruiting medication-free volunteers [22]. The impact of PPs intake on QoL in Post-M women was defined as a secondary outcome. With 80% power, 95% confidence, and an anticipated 15% dropout rate, a minimum of 40 completers was considered necessary to provide an initial evaluation of QoL changes, as reported elsewhere [39].

Three days before and during the intervention, including the washout period, Post-M participants consumed a low-polyphenol diet supervised by a nutritionist and based on pasta, grilled meat or fish, low-fat cheese, rice, and bread, with a limited contribution (1 portion/day) of fruits and vegetables, legumes, nuts and seeds, juices, olive oil, coffee, and tea. Participants were provided with several printed low-polyphenol menus; the study nutritionist contacted participants regularly, and 3-day dietary records were collected and reviewed at each visit to monitor adherence. Volunteers completed a questionnaire developed by the Spanish Association for the Study of Menopause (AEEM), which includes items on educational level, relationship status, employment situation, sexual activity, and alcohol and tobacco consumption [6].

Study 2 aimed to (i) assess the distribution of polyphenol metabotypes and MCs and compare them with those of Pre-M women; (ii) evaluate the impact of chronic polyphenol consumption on these distributions; and (iii) explore the effect of PPs on the QoL of Post-M women, as assessed by the C-SF scale, following 8 weeks of PP intake compared to baseline and placebo.

After providing informed consent, Post-M participants consumed three capsules of PPs daily in the evening for three consecutive days. This intake was used to determine the baseline distribution of metabotypes and MCs, analogous to the metabotyping performed in Pre-M participants. Baseline clinical QoL assessments were performed prior to any supplementation, immediately after informed consent was obtained. At that same visit, participants received the capsules for the 3-day metabotyping intake, which was completed before randomization. This distinction ensures that QoL baseline values reflect participants' unsupplemented status.

Study 2 employed a crossover design (Figure 1), which allowed for within-subject comparisons across conditions, minimized inter-individual variability, enhanced statistical power, and reduced the sample size required to detect significant effects. Post-M participants were randomized in a 1:1 ratio using computer-generated numbers, with constraints to ensure equal allocation to each treatment sequence. No blocking or stratification was applied. Participants consumed either three capsules of PPs daily in the evening or the placebo (Pla) for 8 weeks. Capsule intake ceased at the end of the first period, followed by a one-month washout, during which participants continued the controlled diet. They then crossed over to the second period, switching treatments so that those who had received PPs now received Pla, and vice versa, again for 8 weeks. Participants were instructed to return any remaining capsules after each study period. Adherence was calculated as the percentage of capsules consumed relative to the expected intake. The total follow-up lasted 5 months. At each sampling point (T), participants provided a urine sample, completed the C-SF, and reported any adverse events or protocol-related incidents (Figure 1).

Participants completed the abbreviated version of the Cervantes scale (C-SF) at the four sampling points of the trial (Figure 1A). The C-SF comprises 16 items distributed across four dimensions: menopause and health (with nine subdomains, including vasomotor symptoms, general health, and aging), psychology (three subdomains: anxiety, depression, and fatigue), sexuality (two subdomains: importance of sex and satisfaction with sexual life), and couple relationship (two subdomains: happiness in the relationship and perceived role within it). Each item is scored from 0 (best) to 5 (worst). Higher overall scores indicate greater impact of menopausal symptoms and, consequently, poorer QoL. The collected score was entered into the AEEM calculator, a validated tool that measures postmenopausal QoL from 0 (no impact) to 100 (maximum impact). The C-SF version captures the same health constructs as the larger Cervantes scale but offers greater ease of use in clinical practice [5].

A change of 6.7 points in the global score has been proposed as the minimal clinically important difference (MID) for the C-SF, based on its association with functional outcomes, including sleep quality, productivity, and healthcare utilization [6]. Although no MID thresholds have been formally established for individual domains, proportional estimates based on item distribution suggest that MID may correspond to approximately 3.1 points for the menopause and health domain and 1.6 points for the psychological domain. All these values were used as reference thresholds to interpret the magnitude of domain-specific changes observed in the present study.

Urine samples were centrifuged at 14,000× g for 10 min, filtered through a 0.22 μm PVDF filter, and diluted 1:10 with Mili-Q water containing 0.1% formic acid. Hesperetin (5 μM) was used as an internal standard. It was added to each sample after extraction before analysis. As previously reported [36], diuresis was standardized by measuring urinary creatinine excretion.

Analyses were performed via UPLC-QTOF-MS using an Agilent 1290 Infinity UPLC system coupled to a 6550 accurate-mass quadrupole-time-of-flight (QTOF) mass spectrometer (Agilent Technologies, Waldbronn, Germany) equipped with an electrospray interface (Jet Stream Technology; Agilent Technologies). Separation was carried out using an InfinityLab Poroshell 120 EC-C18 column (3 × 100 mm, 2.7 μm; Agilent Technologies) operating at 30 °C. The mobile phase consisted of water (solvent A) and acetonitrile (solvent B), both of which contained 0.1% formic acid. A flow rate of 0.4 mL/min was applied, and the elution gradient was set as follows: 0–3 min, 5–15% B; 3–11 min, 15–30% B; 11–14 min, 30–50% B; 14–16 min, 50–90% B; 16–18, 90% B; 18–20 min, 95–5% B. Finally, an isocratic step of 1 min duration at 5% solvent B was performed, resulting in a total separation time of 21 min. The ESI operated with the parameters reported elsewhere [36]. Data were processed using Mass Hunter Qualitative Analysis software (version B.08.00, Agilent Technologies, Waldbronn, Germany).

Metabolites were identified by direct comparison with available analytical standards whenever possible. Spectral properties, including molecular mass and fragmentation patterns, were also evaluated to confirm assignments. Tentative identification was performed using the same criteria when standards were unavailable. Additionally, extracted ion chromatograms (EICs) with a narrow mass extraction window (0.01 m/z) were used for area calculation and quantification, thereby minimizing the risk of misinterpreting overlapping peaks [36].

All metabolites used to classify participants into the different metabotypes, i.e., equol producers (EP), urolithin A producers (UMA), and IsoUrolithin A and urolithin B producers (UMB), were quantified using validated UPLC-ESI-qTOF-MS methods developed and reported previously under our analytical conditions [36,40]. These validated methods (calibration curves, limits of detection and quantification, repeatability, and recoveries) ensured accurate identification and quantification of phase II metabolites (sulfates and glucuronides) in biological matrices.

For other metabolites lacking commercial standards (e.g., isourolithin A sulfate), LOD and LOQ values could not be experimentally established. These compounds were therefore tentatively identified based on their accurate mass, isotopic pattern distribution, and molecular formula and were included in the analysis only when the signal-to-noise ratio (S/N) exceeded 10, ensuring confident detection and analytical reliability. The identification of lunularin conjugates was performed following the same analytical criteria and methodology previously described by our group [25], which allowed for consistent characterization of lunularin producers and non-producers in human studies.

The impact of PPs intake on C-SF scores was evaluated using a two-way repeated-measures ANOVA model, assessing intra- and inter-group changes in the dependent variable across time and treatment conditions (SigmaPlot v. 16.0, Systat Software, San Jose, CA, USA; the jamovi project 2025, jamovi v. 2.6.26, retrieved from https://www.jamovi.org↗, accessed on 12 July 2025; Sydney, Australia). This model enabled us to assess the main effects of each factor and potential interactions among them. The Shapiro–Wilk test was applied to examine the normality of the data. When significant differences were detected, post hoc multiple comparisons were performed using the Tukey test to identify specific contrasts between groups.

Analyses were conducted both by considering all participants as a single group and by stratifying them according to individual metabotypes and their clusters (MCs). Changes in C-SF scores were based on least-squares means (LSM) derived from a two-factor repeated-measures Analysis of Variance (ANOVA), which accounted for within-subject variability and the study's crossover design. Degrees of freedom were adjusted using the Greenhouse–Geisser correction when the assumption of sphericity was violated, as indicated by non-integer residual degrees of freedom. This approach provides adjusted estimates for each condition, enabling robust inferences regarding treatment effects.

To explore the potential influence of covariates, changes in C-SF scores (post-treatment vs. baseline) were correlated with BMI, chronological age, age at menopause onset, years since menopause, and PREDIMED score, using multiple linear regression models or Pearson or Spearman coefficients, depending on data distribution. Comparisons between two independent groups (e.g., Pre-M vs. Post-M at baseline) were performed using the independent t-test for normally distributed data or the Mann–Whitney U test for non-normal data. When comparing two dependent groups (e.g., before vs. after PPs or placebo within each arm during both trial periods, i.e., before and after washout and crossover), paired Student's t-tests were used for normally distributed data and the Wilcoxon signed-rank test for non-normal data.

Participants were classified into low, medium, and high producers using unsupervised K-means clustering (k = 3) based on creatinine-normalized metabolite intensities. Data were log-transformed and autoscaled prior to analysis in MetaboAnalyst 5.0 (https://www.metaboanalyst.ca↗, accessed on 15 September 2025). Clustering was performed using the Hartigan–Wong algorithm with Euclidean distance. The thresholds defining the three clusters correspond to the centroid values automatically determined by the algorithm. For each group, the sum of the specific metabolites was considered to define the production groups as follows: equol producers (EP), including equol glucuronide and sulfate isomers; lunularin producers (LP), including lunularin glucuronide and sulfate isomers; urolithin A producers (UMA), including urolithin A glucuronide and urolithin A sulfate; and urolithin B producers (UMB), including isourolithin A glucuronide and sulfate as well as urolithin B glucuronide and sulfate. For visualization purposes, the clusters were represented using non-log-transformed values to reflect the actual relative metabolite concentrations in each group. Urine samples were collected three days after polyphenol mixture intake in Pre-M women. In Post-M women, samples were collected three days after intake, following the baseline visit (T1 + 3d), and again after eight weeks of supplementation. Data plots were performed using GraphPad Prism 9.1.1 software (GraphPad Software, San Diego, CA, USA) and Sigma Plot 16.0. Statistical significance was set at * p < 0.05.

Table 2 shows the main participants' characteristics at baseline. In Pre-M, 160 eligible participants were contacted, and 120 agreed to participate (Study 1). Among Post-M participants, 185 were contacted, and 90 were enrolled in Study 2 (Table 2; Figure 1).

The distribution of individual metabotypes was broadly similar between Pre-M and Post-M women at baseline (Table 2; Figure 2). For urolithin metabotypes, UMA was the most prevalent in both groups, with a slight increase from 64.7% in Pre-M women to 71.9% in Post-M women. UMB showed a decrease (32.8% to 28.1%), and UM0 was detected at a low frequency (2.5%) in Pre-M and was absent in Post-M. EP and ENP showed nearly identical distributions across groups (≈50%). Finally, LP decreased from 60.5% in Pre-M to 50.6% in Post-M. Taken together, these findings indicate that menopausal status exerts only a minor influence on the distribution of individual metabotypes in the studied population.

The distribution of MCs showed some differences between Pre-M and Post-M women (Table 2; Figure 3). In Pre-M, MC3 was the most prevalent (21.8%), followed by MC5 (18.5%), MC2 (14.3%), and MC4 (12.6%). MC1 and MC7 were less frequent (10.9% and 10.1%, respectively), whereas MC6, MC8, MC9, MC10, and MC11 showed very low prevalence (<6%), and MC12 was not detected. In Post-M, MC7 became the most prevalent cluster (22.5%), followed by MC5 (18.0%) and MC2 and MC3 (15.7% each). MC1 and MC4 had intermediate prevalence (10.1% and 9.0%, respectively), while MC6 and MC8 were infrequent (<6%). Finally, MC9–MC12 were absent.

Overall, MC5 and MC2 were among the more common clusters in both groups. The main difference was a shift in dominance from MC3 in Pre-M to MC7 in Post-M, along with the disappearance of MC9–MC11 after menopause. These findings suggest only moderate changes in MC distribution associated with menopausal status.

Regarding Study 2 (Figure 1), 78 Post-M participants completed the 5-month follow-up trial. The adherence to the low-polyphenol regimen was high, and 100% of scheduled 3-day dietary records were returned and reviewed. Compliance with the capsule-intake protocol was nearly complete; only a few volunteers returned capsules corresponding to several days, all from the start of the first period. A few volunteers reported some minor side effects after taking the polyphenol-rich mixture (PPs), including mild gastrointestinal discomfort and gas. However, these effects were transient and did not result in any dropouts from the study.

After 8 weeks of PPs intake in Post-M, some shifts in the distribution of individual metabotypes were observed (Figure 2). UMB increased from 28.1% at baseline to 37%, whereas UMA consequently decreased from 71.9% to 63%, and UM0 remained absent. EP increased slightly (from 50.6% to 59%), with a corresponding decline in ENP (from 49.4% to 41%). LP also increased compared with baseline (56% vs. 50.6%), reversing the nearly balanced distribution observed before the intervention. Again, these changes suggest a modest rebalancing of urolithin, equol, and lunularin metabotypes with PPs intake. However, the overall distribution remained broadly similar to both baseline values and those observed in Pre-M women.

Regarding metabotype clusters (MCs), the most prevalent after intervention were MC3 (20.5%) and MC7 (19.2%), followed by MC1 (12.8%), MC4 (14.1%), and MC5 (14.1%) (Figure 3). The remaining, less prevalent clusters, grouped as "Other," accounted for 10.3%. Therefore, compared with the baseline in Post-M, PPs intake decreased slightly in MC7 (22.5% to 19.2%), whereas MC3 increased (15.7% to 20.5%). MC1 and MC4 both showed higher frequencies than at baseline (10.1% to 12.8% and 9.0% to 14.1%, respectively), while MC2 declined (from 15.7% to 9.0%) (Figure 3). In comparison with the Pre-M group, the distribution of MCs in Post-M after consuming PPs still differed in the relative prominence of MC7 over MC3. However, the gap narrowed due to the increase in MC3 with PPs intake (Table 2, Figure 3). Overall, PPs intake led to moderate shifts in individual metabotypes and MCs in Post-M women, with partial convergence toward Pre-M patterns, but no significant reconfiguration of metabotypes and MCs prevalence.

Following the intake of 3 capsules of PPs per day for three consecutive days, no significant differences (p > 0.05) were observed using the t-test or Mann–Whitney–Wilcoxon tests in the quantitative production of metabolites derived from Uro-A (produced in UMA), Uro-B and IsoUro-A (UMB), equol (EP), and LUNU (LP) between Pre-M and Post-M women at baseline. Production of Uro-A-derived metabolites was 1.38-fold higher in Pre-M compared to Post-M (p = 0.417). For Uro-B-derived metabolites, production was 1.15-fold higher in Post-M versus Pre-M (p = 0.656). Equol production was 1.2-fold higher in Pre-M women (p = 0.216), and lunularin-derived metabolites were 1.3-fold higher in Pre-M versus Post-M (p = 0.688). Therefore, menopausal status per se did not significantly affect the quantitative production of these metabolites.

Next, we examined the distribution of low, medium, and high producers for these metabolites between Pre-M and Post-M women at baseline (Figure 4). For this analysis, representative conjugated metabolites from each metabotype were included: for UMA, Uro-A glucuronide and Uro-A sulfate; for UMB, IsoUro-A glucuronide, IsoUro-A sulfate, Uro-B glucuronide, and Uro-B sulfate; for EP, equol glucuronide and equol sulfate isomers; and for LP, LUNU glucuronide and LUNU sulfate isomers. A higher percentage of low producers was observed in the Post-M group: 5% more for metabolites derived from UMB and equol, 10% more for UMA derivatives, and 3% more for LP. A lower percentage of high producers was also observed for each metabolite class, with a particularly notable 8% decrease in LP high producers in Post-M compared to Pre-M.

No significant differences in metabolite production were observed in Post-M women after PP consumption for Uro-A, Uro-B, and lunularin metabolites when analyzed using the Mann–Whitney U test. Production of Uro-A-derived metabolites was 1.3-fold higher after PP intake (p = 0.966). In the case of Uro-B-derived metabolites, production was 1.6-fold lower after PP consumption (p = 0.768). Finally, production of lunularin-derived metabolites was 1.1-fold lower following PP intake in Post-M women (p = 0.441). However, in EP, the median equol production increased 2.8-fold (mean of 13.6 ± 5.5 SEM) following PPs intake, with 86% of EP showing an increase in equol concentration by the end of the intervention.

Regarding the distribution of low, medium, and high producers of metabolites after 8 weeks of PPs intake in Post-M women, no apparent shift was observed, contrary to expectations, from low to higher producers (Figure 4). In contrast, the proportion of low producers remained stable or even increased for LUNU derivatives. Only in EP was a trend toward increased high producers observed, while the percentage of low producers remained unchanged.

PPs supplementation over 8 weeks resulted in statistically significant improvements in QoL scores on the C-SF compared with placebo (Pla), both in the global score and the menopause and health domain. In the full sample (n = 78), the effects of PPs intake were statistically significant but not clinically relevant (score reduction of −6.7 or lower) compared with placebo for both the global (p = 0.023) and menopause and health domains (p < 0.001) (Table 3). Within the PPs group, all domains showed significant (p < 0.05) pre–post improvements, except for sexuality and couple relationship (Table 3). More importantly, clinically relevant changes (equal to or above the MID, i.e., score reduction ≥ −6.7 points) were observed in specific metabotypes and MCs, such as EP, UMB, MC3, and MC4, with improvements ranging from 20% (UMB) to 41.3% (MC4) in the Menopause and Health domain. Although the clinically meaningful change for a subdomain such as hot flashes has not been established, the robust and significant reduction (−10.4 LSM, PPs vs. baseline; p = 0.001), together with the 95% CI not crossing zero, and a considerable effect size (Cohen's d = −3.39 for PPs vs. placebo), strongly suggests that the improvement observed in the Health domain was mainly driven by the marked decrease in hot flashes incidence among EP volunteers. In the Psychic domain, a clinically relevant change was detected when considering all participants, specifically in the EP metabotype and MC3, both of which are equol producers (Table 3). Thus, although this relevant effect was observed in the overall group, it was primarily driven by participants who produce equol. These results suggest that the impact of PPs was statistically detectable and, more importantly, clinically meaningful, particularly in the EP metabotype within the domains of menopause and health and psychic well-being. Specifically, in EP participants (n = 46), 78% showed a reduction in the Menopause and Health domain score after PPs intake, and 63% achieved a clinically meaningful improvement (≥6.7-point reduction), with a 95% confidence interval of 48.7–75.7%. In the Psychic domain, 71% of EP participants improved, and 58% reached the MID threshold, with a 95% confidence interval of 43.3–71.6%.

The participants' diet did not influence the observed effects, as adherence to the low-polyphenol diet was high, and all reported a strong adherence to the Mediterranean diet in their usual dietary habits. Multiple linear regression models were used to assess the impact of diet while controlling for BMI, age, and ethnicity (98% Caucasians), confirming that these covariates did not significantly affect the outcomes. It is essential to highlight that no carry-over effect was observed. Figure 5 illustrates differences observed in the Menopause and Health and Psychic domains, comparing ENP and EP metabotypes. After PP intake, scores do not remain low but return to baseline levels, indicating no carry-over effect. A defined response is observed in EP but not in ENP, and no significant difference is seen between PPs and placebo.