Influence of dietary components on the gut microbiota of middle-aged adults: the gut-Mediterranean connection

Abstract

Background

A plant-focused, healthy dietary pattern, such as the Mediterranean diet enriched with dietary fiber, polyphenols, and polyunsaturated fats, is well known to positively influence the gut microbiota. Conversely, a processed diet high in saturated fats and sugars negatively impacts gut diversity, potentially leading to weight gain, insulin resistance, and chronic, low-grade inflammation. Despite this understanding, the mechanisms by which the Mediterranean diet impacts the gut microbiota and its associated health benefits remain unclear.

Methods

This retrospective, observational study explored the relationships between Mediterranean dietary components—vegetables, fruits and nuts, legumes, whole grains, fish, meat, dairy, alcohol, saturated and unsaturated fats—and the gut microbiota in middle-aged adults enrolled in Alberta’s Tomorrow Project, Canada. Diet was recorded using the Canadian Dietary History Questionnaire (CDHQ-II) and participants were classified into four quartiles based on a modified Mediterranean Diet Score. Blood and fecal samples were collected for metabolomics and 16S rRNA sequencing, respectively.

Results

Findings revealed that higher adherence to the Mediterranean diet was associated with increased alpha diversity and a greater abundance of beneficial fiber-degrading bacteria, including Prevotella, Parabacteroides, Clostridium XIVb, Coprobacter, and Turicibacter. Furthermore, participants who consumed more Mediterranean diet components exhibited higher concentrations of serum microbial metabolites including p-hydroxy hippuric acid and indole-acetaldehyde.

Conclusions

Results demonstrate a pivotal role of the gut microbiota, via its metabolites in harnessing the health benefits of the Mediterranean diet, highlighting its potential to promote metabolic health and prevent chronic disease.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12866-025-04170-6.

Introduction

The Mediterranean diet (MedD), characterized by a higher intake of fruits, vegetables, whole grains, legumes, olive oil, and fish, along with a modest consumption of red meat and saturated fats has been gaining interest as one of the healthiest dietary patterns [1–3]. Several epidemiological and clinical studies have associated adherence to the MedD with increased longevity and a reduced incidence of chronic metabolic, inflammatory, and gastrointestinal disease [2, 4–6]. Despite the well-documented benefits of the MedD, the global rise in obesity and its associated comorbidities [7] suggests that many populations are shifting away from such healthy eating patterns. Instead, the widespread adoption of a high-calorie, low-nutrient diet commonly known as the 'Western diet' has become a primary contributor to this health crisis. Characterized by the extensive availability and consumption of affordable, ultra-processed foods high in sugar, salt, refined grains, and saturated fats, the Western diet is linked to a range of adverse health outcomes, including weight gain, chronic metabolic disease, and an elevated cancer risk [8].

Recent research highlights the significant impact of diet on the gut microbiota, the diverse community of microorganisms residing in the human digestive tract [9–12]. Notably, the MedD, which is rich in fiber, polyphenols, and healthy fats, fosters a gut microbiota composition that is instrumental to delivering its health benefits [3, 13]. In stark contrast, diets high in saturated fats, sugars, and animal proteins are associated with reduced microbial diversity and an increased risk of gut dysbiosis [14, 15]. For instance, a randomized controlled study in adults (18–65 years) at risk of obesity found that increased adherence to the MedD was associated with a higher abundance of the Bacteroidetes phylum and a decrease in genera from the Firmicutes phylum as well as a reduced body weight and fat mass in participants [16]. Likewise, transition from a Western diet to a MedD also imparts benefits. Just four days of the MedD led to a significant increase in beneficial fibrolytic taxa such as Lachnospiraceae and Butyricicoccus bacteria as well as gut-derived tryptophan metabolites in young adults (18–25 years) [17]. Similarly, a 12-month MedD intervention increased the abundance of carbohydrate-degrading Faecalibacterium prausnitzii and Roseburia species and enhanced the production of short- and branched- chain fatty acids in elderly subjects (> 65 years), regardless of age or body mass index (BMI) [18].

Despite evidence supporting the benefits of a plant-focused, MedD diet, the mechanisms by which it confers health benefits are not yet fully elucidated. One proposed pathway involves the interaction between gut microbiota, dietary fiber, and phenolic compounds present in the MedD, which together produce distinct, bioactive metabolites that improve cardiometabolic health [13]. The present study examined relationships between MedD components and the gut microbiota in middle-aged adults (40–65 years), while controlling for potential confounding factors including age, BMI, and sex. Additionally, the study investigated microbial metabolites associated with the MedD, providing insights into how specific diet-driven modifications in gut microbiota contribute to overall health.

Methods

Study participants

This retrospective, observational study was approved by the Conjoint Health Research Ethics Board at the University of Calgary (REB17-1973). Alberta’s Tomorrow Project is a population-based cohort and has approximately 55,000 individuals enrolled between 2000 to 2015 [19, 20]. A subset of 1343 participants, still living in Calgary (Alberta, Canada) were recontacted via phone or email for participation in the study. In all, 443 adults (28.2% males and 71.8% females) between the age of 38–65 years were recruited based on their age, BMI, medical history and health status. Of these, 75 were excluded based on pre-determined inclusion criteria that included valid questionnaire completion, fecal and blood collection, absence of Class III obesity (BMI > 40 kg/m²) as well as pregnant women, cancer patients, heavy smokers, and/or individuals who used antibiotics in the past 3 months. A CONSORT flowchart is shown in Supplementary Fig. 1.

Once written informed consent was obtained, the Canadian Dietary Health Questionnaire (CDHQ II) was sent to eligible participants via email. A modified Mediterranean diet score (mMDS) was calculated based on the weekly intake of nine food groups adjusted by total energy intake– 1) Vegetables, 2) Fruits, 3) Legumes, 4) Grains, 5) Dairy, 6) Meat, 7) Fish, 8) Alcohol, and 9) Fatty acid ratio (calculated as the sum of mono- and poly-unsaturated fatty acid divided by saturated fatty acid intake). Sex-specific median values for healthy populations were used as cut-offs for each food component. The diet scores ranged from 0 (least healthy) to 9 (most healthy). No participants had scores of 0 or 9, resulting in four mMDS-based quartiles: Q1 (mMDS 1 and 2, n = 64), Q2 (mMDS 3 and 4, n = 140), Q3 (mMDS 5 and 6, n = 128), and Q4 (mMDS 7 and 8, n = 36). A detailed breakdown of dietary intake of macro- and micronutrients is shown in Supplementary Table 1.

Sample collection, and anthropometric measurements

A fecal collection kit (Protocult 120, Ability Built, USA) was mailed with instructions to collect samples at home. Participants were informed to store the collected stool sample at −20 °C and bring the sample to the study site (Richmond Road Diagnostic and Treatment Center, Calgary, Alberta, Canada) on the day of the appointment. Upon arrival, a trained phlebotomist obtained blood via the antecubital vein and isolated serum. Fasting blood and fecal samples were collected within 48 h. Collected samples were then stored at −80 °C until further processing. During the appointment, body measurements such as standing/sitting height, weight, systolic/diastolic blood pressure, heart rate, waist-hip circumference, and hand grip strength were also recorded by the trained staff at the study site. Details of these procedures have been previously published [21].

Fecal DNA extraction, amplification, and sequencing

Detailed steps on sequencing, quality control and rarefaction have been previously described [21, 22]. Briefly, total genomic DNA was extracted from fecal samples using the QIAamp Fast DNA stool mini kit (Qiagen, Germany). Sample (~ 200 mg) was mixed with 1 ml InhibitEX buffer in 0.70 mm Garnet bead tubes, homogenized using a SpeedMill PLUS (Analytic Jena, Germany) at 60 Hz (45 s), heated at 95 °C (5 min) and finally centrifuged at 10,000 rpm (60 s). The obtained supernatant (200 µl) was transferred to a 1.5 ml microcentrifuge tube and then placed in the QIAcube for follow-up automated DNA isolation according to the manufacturer’s protocol. The purified DNA pellet was resuspended in 200 µl of TE buffer and stored at −20 °C until further use. Samples were sequenced following amplification of the V3 − V4 hypervariable region. Demultiplexed sequences generated for this study are deposited on the NCBI sequence read archive (SRA PRJNA922681). Preprocessing of the 16S rRNA data was conducted in R (v.4.1.3) using phyloseq package (v.1.34.0). Rare ASVs such as those with less than 20 sequencing reads, ASVs belonging to Cyanobacteria phylum, mitochondria family, and Chloroplast class were removed giving a total count of 18,937 ASVs.

Untargeted serum metabolomics

Briefly, 50 µl of each serum sample was mixed with 200 µl of ice-cold, LC–MS grade 100% methanol (5 µM) containing internal standard, sonicated, and centrifuged at 14,000 rpm (10 min). The supernatant was collected and placed in the Speed-Vac system (Eppendorf Vacufage, Enfield, USA) at 45 °C until completely dry. Extracts were then reconstituted using 50% methanol, centrifuged, and filtered twice using a 0.2 µM filter to remove any impurities. Samples were analyzed using Agilent 6550 iFunnel Q-TOF LC/MS system (Agilent, Santa Clara, CA, USA). Metabolites were separated using Acquity UPLC HSS T3 C18 column (2.1 × 100 mm, 1.8um size; Waters Corporation, USA) in gradient mode (Mobile Phase H₂O, Mobile Phase 100% Acetonitrile). Data processing and machine parameters are described in detail elsewhere [23]. To check the reproducibility, quality control samples were run after every 12 samples. Raw data were converted using ProteoWizard 4.0 software to mzXML format and uploaded to XCMS online (https://xcmsonline.scripps.edu/) [24]. Data were normalized using median-fold change and auto-scaling features in MetaboAnalyst 5.0 (https://www.metaboanalyst.ca/) for the extracted peak abundance [25]. Features were then identified using the Human Metabolome Database, HMDB (http://www.hmdb.ca/) [26], PubChem (https://pubchem.ncbi.nlm.nih.gov) [27], and METLIN databases (http://metlin.scripps.edu) [28]. For the volcano plot, we defined significance based on two criteria: fold change threshold and p-value threshold. A log2 fold change greater than 1 (or less than −1), corresponding to a minimum two-fold change in expression and a p-value < 0.05, which corresponds to -log10 (p-value) > 1.3.

Statistical analysis

All data are presented as mean ± SD unless otherwise specified. An unpaired t-test with Welch’s correction was performed for comparisons between two groups while a one-way ANOVA was performed for multi-group analyses. Where applicable, outliers were identified and removed using the ROUT [29] set at 1% in GraphPad Prism v.10 (Boston, USA). A p ≤ 0.05 was considered significant. Relative microbial abundances were center log-ratio (CLR) transformed following zero-replacement using zCompositions (v.1.3.4) and CoDaSeq (v.0.99.6) to control for composition before statistically assessing differential abundance. Alpha and beta diversity were calculated using the Shannon Diversity Index [30] and Bray–Curtis dissimilarity metric [31].

Differences in alpha and beta diversity across the study groups were assessed using Kruskal–Wallis and Permutational Multivariate Analysis of Variance (PERMANOVA, ‘adonis2’ package in R, v.5.4.1), respectively. Linear Discriminant Analysis Effect Size (LefSe) was performed for the identification of biomarkers using factorial Kruskal–Wallis with a cut-off of p < 0.05 (https://huttenhower.sph.harvard.edu/galaxy/). The association between the microbiota and its covariates (BMI, age, sex, and blood pressure) was measured using redundancy analysis [32]. To further assess taxonomic associations, multivariable modelling between the microbial features and their covariates (BMI, age, and sex), as well as all food components were examined by Multivariable Association with Linear Models (MaAsLin2; R package ‘Maaslin2’, v.1.41.1) [33]. Results were corrected for multiple testing using the false discovery rate (FDR); associations with q ≤ 0.25 were considered significant.

Results

Participant characteristics

Study participant characteristics are shown in Table 1. The study cohort comprised 368 participants (266 females/102 males) between the ages of 40 to 65 years (mean ± SD = 56.9 ± 6.3 years) with an average BMI of 25.2 kg/m². Of these, 65% participants were normal weight, 20% were overweight (25.0–29.9 kg/m²) and 15% were obese (> 30 kg/m²). The mMDS was calculated as previously described [6]. Participants were categorized into quartiles based on their mMDS, which ranged from 0 (least healthy) to 9 (most healthy) (Fig. 1A). Participants in Q3 had a lower BMI compared to those in Q1 with no differences between other quartiles (Fig. 1B). However, when stratified by waist-to-hip ratio, there were no differences between any of the quartiles (Fig. 1B).

Table 1.

Participant characteristics by mediterranean diet quartiles

	Mediterranean diet quartiles
	Q1	Q2	Q3	Q4
Number (F)	64 (49)	140 (99)	128 (95)	36 (23)
Age (years)	56.2 ± 6.7	57.3 ± 5.9	57.1 ± 6.4	56.3 ± 6.2
BMI (kg/m2)	26.5 ± 4.5	25.1 ± 4.4	24.6 ± 4.1*	25.5 ± 4.3
Waist: Hip ratio	0.9 ± 0.1	0.9 ± 0.1	0.9 ± 0.1	0.9 ± 0.1
Weight (kg)	74.4 ± 14.8	72.1 ± 17.5	70.5 ± 14.2	74.4 ± 15.5
SBP (mmHg)	122.1 ± 13.5	122.8 ± 16.9	120.3 ± 15.6	120.5 ± 11.3
DBP (mmHg)	76.3 ± 9.5	75.7 ± 11.4	72.9 ± 9.9	74.3 ± 8.2
Heart rate (bpm)	65.9 ± 9.7	63.9 ± 10.4	64.0 ± 8.8	62.3 ± 9.7
Diet Score	1.7 ± 0.5	3.5 ± 0.5*	5.4 ± 0.5*	7.2 ± 0.4*

Data are presented as mean ± SD (n = 368)

Abbreviations: BMI body mass index, bpm beats per minute, DBP diastolic blood pressure, F female, Q1-4 quartiles 1 to 4, SBP systolic blood pressure

^*Indicates a significant difference from Q1 (p < 0.05)

Fig. 1.

Study participants. A Distribution of study participants into different diet quartiles based on their modified Mediterranean diet score (mMDS). Q1 = 1,2 mMDS; Q2 = 3,4 mMDS; Q3 = 5,6 mMDS; Q4 = 7,8 mMDS. The diet scores ranged from 0 (least healthy) to 9 (most healthy) and was based on weekly consumption of 9 food components adjusted by total energy intake– vegetables, legumes, fruits and nuts, dairy, whole grains, meat, fish, alcohol, and fatty acid ratio. No individuals scored 0 or 9. B BMI and WHR. Abbreviations: BMI; body mass index, mMDS; modified Mediterranean diet score, WHR; waist-to-hip ratio

Mediterranean diet enriches participant microbial diversity and composition

Diversity analyses showed that participants in Q4, with higher adherence to the MedD exhibited greater microbial evenness and richness as reflected in the Shannon index (Q4: 4.34 ± 0.24 vs. Q1: 4.17 ± 0.34, p < 0.05) (Fig. 2A). However, no significant distinctions in beta diversity were observed across the quartiles (Fig. 2B). Similarly, no notable differences were found in the relative abundance between the four quartiles at either the phylum or family level (p > 0.05) (Figs. 2C and D). For example, the relative abundance of Firmicutes was comparable across Q1 (45.0%), Q2 (44.7%), and Q3 (45.4%), with a modest reduction in Q4 (38.8%). Similarly, Bacteroidetes ranged from 39.8% (Q1) to 46.3% (Q4), and Verrucomicrobia remained relatively stable (5.1%–5.9%). Actinobacteria ranged from 1.7% (Q4) to 2.8% (Q1), and Proteobacteria ranged from 5.0% (Q2) to 8.1% (Q4). These differences were not statistically significant.

Fig. 2.

Impact of Mediterranean diet on gut microbial diversity and composition. A microbial richness and evenness (alpha diversity) as estimated by Shannon Index; B variability and distribution of microbiota structure (beta diversity) as determined using principal coordinate analysis across all four diet score quartiles; relative abundance at C phylum and D family level; E–H Linear regression of diet score and genera identified using Microbiome Multivariable Association with Linear Models (MaAsLin2) analysis

Results showed a progressive increase in genera Prevotella (coef = 0.07, p = 0.01) and Turicibacter (coef = 0.16, p < 0.01) and a decrease in Blautia (coef = −0.17, p < 0.01) and Coprococcus (coef = −0.07, p < 0.01) (Fig. 2E-H) with higher adherence to MedD patterns indicating distinct microbial profiles between participants with the lowest (Q1) and highest (Q4) adherence to the diet (p < 0.01).

To delve deeper into the differential microbial composition driving the changes in diversity between Q1 and Q4, a Linear Discriminant Analysis Effect Size (LEfSe) analysis was conducted [34]. The analysis showed that individuals adhering to MedD pattern (Q4) had a higher abundance of Prevotella (LDA score ~ 2.8) and its species P. corporis (~ 1.9), as well as Proteobacteria (~ 3.8), Betaproteobacteria (~ 3.7), and its family Sutterellaceae (~ 3.3). In contrast, those following lower adherence to MedD (Q1) exhibited enrichment of Actinobacteria (~ −2.3) and its family Coriobacteriaceae (~ −2.4), as well as Gemella (~ −3.5), Bacillales (~ −3.9), and its species G. sanguinis (~ −2.8) (Fig. 3A). Detailed figures illustrating the differences in taxa abundance between Q1 and Q4 at the genus and species levels are shown in Supplementary Figs. 2 and 3. At the genus level, MaAsLin2 analysis adjusted for age, BMI and sex confirmed that the Q4 group had a higher abundance of Clostridium XIVb (p < 0.01, q = 0.15), Fusobacterium (p < 0.01, q = 0.15), Coprobacter (p < 0.01, q = 0.20), Turicibacter (p < 0.05, q = 0.27), Prevotella (p < 0.05, q = 0.38) and Parabacteroides (p < 0.05, q = 0.41) when adjusted for age, BMI and sex (Fig. 3B).

Fig. 3.

Differential microbial taxa and their association with diet between Q1 vs. Q4 study groups. A Cladogram diagram showing the microbial taxa with significant differences between the study groups. Microbial classification at the level of phylum, class, order, family, genus, and species is shown from the inside to the outside. The red and green nodes in the phylogenetic tree represent features with significant differential abundance (non-parametric factorial Kruskal–Wallis sum-rank test) between Q1 and Q4 groups, respectively. Yellow nodes represent species with no significant difference; B Microbiome Multivariable Association with Linear Models (MaAsLin2) analysis showing relationship between diet and gut microbial features (diet as fixed effects and age, sex, and body mass index as random effects). *p < 0.05, q < 0.25; #p < 0.05, q > 0.25. Green indicates higher in Q4 vs. Q1, red lower in Q4 vs. Q1

Distinct metabolite patterns are associated with Mediterranean diet consumption

To investigate changes in the serum metabolites influenced by diet, gut microbiota, or other external factors, untargeted serum metabolomics analysis was performed. Using LC–MS/QTOF, over 20,000 features including those related to dietary intake and microbiota composition were assessed. Supervised clustering analysis (PLS-DA) showed distinct separation between the study groups (CV-ANOVA, p < 0.001, Fig. 4A). A volcano plot illustrating the significant increases and decreases in the concentration of metabolites between Q1 and Q4 is shown in Fig. 4B. Healthier eating resulted in an increase in plant-based food metabolites such as betonicine (log₂FC = 2.1), a compound found in citrus fruits, and p-hydroxy hippuric acid (log₂FC = 1.8), a polyphenol-derived microbial metabolite [35]. Indole acetaldehyde (log₂FC = 1.7), produced by the action of gut microbiota on tryptophan-rich foods was also enriched in the higher adherence group [36], alongside phenylalanine betaine (log₂FC = 2.0) and 3-amino-4-hydroxybenzoic acid (log₂FC = 1.6), both linked to fiber- and polyphenol-rich diets. Higher MedD consumption was also associated with declines in fumonisin AK1 (log₂FC = −2.3), a metabolite associated with contaminated foods [37] as well as the food additive polyoxyethylene sorbitan monoleate (P80) (log₂FC =–1.9) [38].

Fig. 4.

Impact of Mediterranean diet on serum metabolomics. A Partial Least Square-Discriminant Analysis between Q1 and Q4 groups; B Volcano plot showing differentially abundant serum metabolites in Q4 vs. Q1 participants. Red dots indicate lower relative peak intensity, and green dots indicate a higher peak intensity. Significance was based on two criteria: i) A log2 fold change greater than 1 (or less than −1), corresponding to a minimum two-fold change in expression and ii) p-value < 0.05, which corresponds to -log10(p-value) > 1.3. Additional significant metabolites as indicated: a. Creatine ethyl ester; b. N-Nitroso(2-hydroxypropyl)(2-oxopropyl)amine; c. 1-(5-Hydroxy-2-oxo-2,3-dihydroimidazol-4-yl)urea; d. N-[4-[(4-Propan-2-yloxyphenyl)methyl]phenyl]−4,5-dihydro-1H-imidazol-2-amine; e. Dimethyl tetrasulfide; f. Nerisopam; g. N-Methyl-N-2-propynyl-1-indanamine; h. Ethyl 4-methyl-5-(1-methylethoxy)−9H-pyrido(3,4-b)indole-3-carboxylate; i. LysoPE(24:1(15Z)/0:0); j. LysoPC(18:2/0:0); k. DG(18:1(12Z)−2OH(9,10)/0:0/i-12:0); l. Not identified; m. DG(22:6(4Z,7Z,11E,13Z,15E,19Z)−2OH(10S,17)/0:0/i-15:0); n. Sambutoxin; o. Noberastine; p. C.I. Food Black 2; q. Theaflavin 3'-O-gallate; r. Tributyrylglycerol

Gut microbiota reflects components of the Mediterranean diet

Before exploring relationships between specific food components and microbial taxa, the validity of the calculated mMDS score and various MedD components was examined. The mMDS exhibited positive correlations with healthier food components such as vegetables, fruits, legumes, fish, whole grains, moderate alcohol, and the fatty acid (FA ratio, calculated as the sum of mono- and poly-unsaturated fatty acid divided by saturated fatty acid intake). Conversely, a negative correlation was observed between mMDS and meat, dairy, and saturated fats (Fig. 5A). Plant-based components such as fruits, vegetables, legumes, and grains displayed an inverse association with meat, dairy, and saturated fats (Fig. 5A).

Fig. 5.

Association between Mediterranean food components and gut microbiome. A Spearman correlation analysis between modified Mediterranean diet score (mMDS) and its dietary components. B Redundancy analysis to estimate explained variation (%) in gut microbial data by dietary groups, and its covariates (age, BMI, sex, PA, and BP); C Spearman correlation analysis of food components and differentially abundant taxa. Abbreviations: BMI, body mass index; BP, blood pressure; fa_ratio, fatty acid ratio, mMDS, modified Mediterranean Diet Score; leg, legume, MUFA, monounsaturated fatty acids; PA, physical activity; PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids, veg, vegetable

Next, a redundancy analysis was conducted to elucidate the proportion of explained vs. unexplained variance in the microbiota data. While the diet quality score could only account for a modest fraction of microbial taxonomy variation, significant (p < 0.01) effects were observed for legumes, vegetables, fruits, and meat (Fig. 5B). Examination of the relationships between gut microbial taxa and food groups showed the consumption of fruits, legumes, whole grains, alcohol, and the FA ratio to be positively correlated with Prevotella and Coprobacter. In addition, Turicibacter showed significant positive associations with fish and saturated fats while Prevotella demonstrated a significant negative association with saturated fats (Fig. 5C).

Discussion

As the incidence of lifestyle-related cardiometabolic disease continues to climb, choosing a health-conscious dietary pattern that promotes beneficial gut microbiota is increasingly important. In this study, comprehensive analysis demonstrated an increased gut microbial diversity likely due to a higher prevalence of carbohydrate-degrading genera such as Prevotella, Clostridium XIVb, Turicibacter, and Coprobacter among individuals who adhere to a healthier dietary regimen. Furthermore, untargeted metabolomics investigation identified elevated concentrations of health-promoting gut-derived metabolites including p-hydroxy hippuric acid, betonicine, and indole acetaldehyde, accentuating the complex interplay between the MedD, gut microbiome, and metabolite production, likely contributing to its health benefits.

This study emphasized a significant impact of increasing MedD consumption on the structural variation of the gut microbiota similar to other studies conducted in large-scale populations in Europe [39] and the USA [40]. Upon categorizing participants by diet quality, a noteworthy trend emerged. Those in the highest quartile (Q4) eating a diet characteristic of MedD exhibited higher alpha diversity, indicative of a potentially lower risk of cardiometabolic disease [41] and enhanced stability and resilience of the gut microbiota [42]. These results are in line with other studies that suggest that a higher intake of fiber-rich foods and a lower intake of red/processed meat might be responsible for the increase in gut microbial diversity [10]. Additionally, results of the current study demonstrate the specific dietary components within the MedD that influence gut microbiota diversity. Among these components, legumes, vegetables, and fruits were identified as primary drivers.

When age, sex, and BMI were accounted for, results showed that increased intake of plant-based, fiber-rich foods likely increased fiber-degrading bacteria. For instance, higher consumption of legumes, vegetables, fruits, nuts, and whole grains that are rich in insoluble fermentable fibers and reduced intake of meat, dairy, and saturated fats were positively linked to beneficial taxa such as Clostridium XIVb, Coprobacter, Turicibacter, Prevotella, and Parabacteroides. These taxa are well-established for their capacity to degrade complex carbohydrates and produce short-chain fatty acids (SCFA), mainly propionate [43, 44]. A recent study by Seethaler et al. showed that a 3-month MedD intervention improved intestinal barrier function by increasing SCFA in women with gut barrier impairment [45]. The authors concluded that the SCFA-mediated dietary effect on the intestinal barrier was mainly due to the consumption of nutrients containing fiber, especially water-insoluble cellulose and fermentable oligosaccharides found in legumes, fruit, and nuts. Furthermore, Merra and colleagues linked a carbohydrate-rich diet with the enterotype dominated by Prevotella and the higher consumption of animal proteins and fats to an enterotype dominated by Bacteroides [10]. Results of the present study confirm a positive association between Prevotella and the intake of legumes, fruits, and the FA ratio, along with a negative association with saturated fats and meat. These findings further support the link between Prevotella and plant-based food components involved in carbohydrate metabolism. Another finding warranting discussion is a positive relationship between Prevotella and alcohol consumption found in this study. Although the types of alcohol consumed were not detailed in this study, red wine consumption is known to be associated with increases in microbiota α-diversity and Prevotella in both intervention and cohort studies [46, 47].

In this study, untargeted serum metabolomics was used to explore changes occurring within the metabolite landscape with differential dietary exposure [48]. Notable elevations in the concentrations of several beneficial microbial-derived bioactive metabolites were found. Among these metabolites, p-hydroxy hippuric acid produced by the action of gut microbiota on dietary polyphenols has garnered substantial attention due to its potential health benefits [35, 49]. This compound has been associated with anti-inflammatory and antioxidant effects, which can contribute to the overall well-being of individuals consuming a MedD rich in polyphenols. Similarly, indole acetaldehyde, a microbial metabolite generated in tryptophan metabolism has been positively linked to gut homeostasis [36]. Interestingly, in this study, the subsequent increase in certain species belonging to Prevotella and Blautia genera, which have been well-reported to be involved in the production of microbial-derived metabolites [35, 44, 50] further highlights the role of metabolites in achieving gut microbiota-mediated, MedD effects.

Additional metabolites of interest included increased concentrations of trimethyltyrosine, phenylalanine betaine, and indoleacetaldehyde in healthy-eating participants, indicative of differential aromatic amino acid (AAA) metabolism. These include tyrosine, phenylalanine, and tryptophan that are primarily obtained from diet and are well-established for their role in protein synthesis [51]. Recently, Liu and colleagues showed that gut microbiota can break down these AAA and generate metabolites that have the potential to influence immune, metabolic, and neuronal responses as signaling molecules [52]. Conversely, metabolites associated with poor dietary habits and the consumption of highly processed foods were also detected. Specifically, polyoxyethylene sorbitan monoleate (P80), a widely used emulsifier in food and pharmaceutical formulations was higher in Q1. Relevant to the present study, this additive has been linked to metabolic disturbance, microbial gut dysbiosis as well as accelerated cognitive decline [53–55].

Limitations of this study include self-reported dietary data which may introduce inherent biases and inaccuracies, potentially affecting the precision of dietary assessments [56]. However, the observed associations between the calculated mMDS and MedD components served to validate the dietary quality among participants. The 16S rRNA sequencing method also limits taxonomic resolution, making it difficult to identify microbes at the species level [57]. The absence of fecal metabolomics analysis further restricts a comprehensive understanding of the metabolic pathways and specific metabolites influenced by the MedD in relation to the gut microbiota. Lastly, due to the study's observational nature, establishing causation between the MedD and alterations in gut microbiota composition or function remains challenging, as it does not account for other unmeasured variables. Still, the use of robust statistical models that account for diet, microbiota, and related metabolites strengthens confidence in the findings.

In conclusion, as lifestyle-related cardiometabolic diseases continue to rise, adopting a diet that fosters beneficial gut microbiota composition is critical. The findings of this study underscore the significant benefits of the MedD on enhancing gut microbiota diversity and improving metabolic health in middle-aged adults. Increased intake of fiber-rich foods like legumes, vegetables, and fruits was associated with a higher abundance of beneficial bacteria such as Prevotella and Clostridium XIVb, and health-promoting metabolites like p-hydroxy hippuric acid and indole acetaldehyde, linking the diet to its well-established anti-inflammatory and antioxidant effects.

Authors’ contributions

ML and JS secured funding to support the execution of the study. GS coordinated the recruitment process and oversaw the collection of biological samples, along with the acquisition of all demographic and survey-related data. NR and KS carried out the 16S rRNA gene sequencing and preliminary processing. SS conducted the metabolomics experiments, developed the overarching study concept, and played a central role in organizing the dataset. SS also performed all statistical analyses, interpreted the results, and created the data visualizations. SS drafted the initial manuscript, with editorial and structural input from CM. Both CM and JS critically reviewed and revised the manuscript for intellectual content. All authors reviewed and approved the final version of the manuscript for submission.

Funding

This work was supported by a Bundesministerium für Bildung und Forschung (BMBF, Germany, FKZ: 0315540); ERA-HDHL Initiative: Gut Metabotypes as Biomarkers for Nutrition and Health (BLE, Germany, FKZ: 2816ERA14E, 2816ERA13E) and the National Science and Engineering Research Council of Canada (NSERC). CM and SS received stipend support from University of Calgary Eye’s High Fellowships.

Data availability

The 16S rRNA gene-sequences data supporting the conclusions of this article are deposited at the National Center for Biotechnology Information (BioProject SRA PRJNA922681).

Declarations

Ethics approval and consent to participate

All participants gave written informed consent in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki). The study was approved by the Conjoint Health Research Ethics Board at the University of Calgary (REB17-1973).

Consent for pubication

All authors have consented to publication. Authors revised the manuscript before final approval and agreed to be accountable for all aspects of the work. All persons designated as authors qualify for authorship.

Competing interests

The authors declare no competing interests.

Microbiota Nomenclature

We acknowledge that recent versions of SILVA and GTDB have updated taxonomic nomenclature, but we retained the legacy terms assigned by the database used at the time of analysis to preserve consistency.