Introduction

Gut microbiota is a complex dynamic microbial system helping in better gastrointestinal function [1]. Although the microbiota concept is not yet fully understood, it is known that millions of bacteria colonized the human intestines, contributing to its formation. The gut microbiota composition extensively affects human health, and each individual's dietary intake plays a major role in the microbiota composition. The interactions between diet and gut microbiota are observed to be mutual [2]. Growing evidence shows that the gut microbiota composition extensively affects the host's immunological, nutritional, and metabolic functions and plays a critical, symbiotic role in human health [3,4,5].

It is observed that the dietary pattern could have a notable role in shaping the gut microbiota composition by providing substrates that can differentially promote the growth of specific microbes and communities [6]. Mediterranean diet (MD) is one of the most studied healthy dietary patterns, characterized by high amounts of fruits, vegetables, nuts, seeds, olive oil, and unrefined grains, moderate quantities of fish, a small amount of poultry, and least possible consumption of red and processed meats [7].

Evidence from the literature illustrates a beneficial effect of MD on metabolic and chronic diseases, including obesity, type-2 diabetic mellitus, cardiovascular disease, and metabolic syndrome, which may be partly through beneficial changes in gut microbiota composition and function [8,9,10]. A high proportion of plant-based foods in MD correlates with a higher percentage of short-chain fatty acids (SCFAs) and fiber-degrading bacteria in the feces [11]. It has been shown that subjects with higher adherence to MD had a lower presence of E. coli and an increased total abundance of bacteria, a higher Bifidobacteria to E. coli ratio, and an increased prevalence of C. Albicans [1].

Bacteria ferment dietary fiber in the colon to produce SCFA, which is believed to have systemic anti-inflammatory effects [12]. Moreover, polyphenols in MD are known to have prebiotic actions that can change gut microbiota and produce metabolites with consequential effects on host health [13]. The effects of MD on the gut microbiota composition have been widely investigated in different studies, but evidence from individual studies is somehow inconsistent. In some studies, higher adherence to MD has resulted in positive gut microbiota diversity [14, 15], yet some evidence reported no change or even decrease of some beneficial bacterial phyla after MD intervention [16, 17].

A wide range of studies targeting different populations have been conducted in this regard; thus, defining the appropriate criteria and summarizing the findings can be challenging. Hence, we conducted this systematic review to investigate the results of observational studies and clinical trials regarding the possible changes in the gut microbiota diversity and abundance, its metabolites, and finally, participants' clinical outcomes following adherence to MD in healthy populations or patients suffering from metabolic disorders.

Methods and materials

Search strategy and selection of studies

A systematic literature search was conducted on PubMed, Web of Science, and Scopus databases. All related articles published up to October 2023 were considered for inclusion. Besides, Google Scholar and recent review articles' references were checked for further article inclusion. Search queries were as following: ("Mediterranean diet"[Title/Abstract] OR "Mediterranean dietary pattern"[Title/Abstract] OR "Mediterranean dietary intervention"[Title/Abstract] OR "Mediterranean-style diet"[Title/Abstract]) AND ("microbiota"[Title/Abstract] OR "microbiome"[Title/Abstract] OR "microflora"[Title/Abstract] OR "microbial profile"[Title/Abstract] OR "microbial composition"[Title/Abstract] OR "bacterial load"[Title/Abstract]).

The method of presenting the topics, including analysis and interpretation, determining the study's objectives, and collecting the findings, was performed based on the preferred reporting items for systematic reviews and meta-analyses (PRISMA) [18].

Eligibility criteria

Two researchers separately screened the titles, abstracts, and then full-text of the articles and selected the relevant studies based on their relevance to the objectives of the systematic review, separately. Disagreements between the two researchers were resolved by discussion or consulting with a third reviewer until reaching a consensus. Duplicate papers retrieved from different queries were removed, and only articles with more complete data were considered. Studies were excluded if the main text was not available or was not in English, if the articles did not investigate the gut microbiota composition following adherence to MD, or if a dietary intervention was not described as MD by the article's authors.

Articles conducted on healthy subjects or patients with metabolic disorders were included in our study. Patients with an inflammatory disease, including Inflammatory Bowel Disease (IBD) and Rheumatoid Arthritis (RA), were excluded from our study. Gut microbiota alterations with/without clinical change following MD were considered outcomes. Reviews, protocols, editorials, letters, case reports, and experimental or animal studies were excluded. Therefore, only observational and interventional studies with original data on humans were included in the present study.

Data extraction

The extraction checklist for both observational and interventional studies consisted of the following parts: Surname of the first author, publication year, country, information on the study design, participants' characteristics (age, gender, and ethnicity), study duration, study cohort, sample size, alpha and beta diversity, microbial alteration, gut microbiota-derived metabolites, and clinical outcomes. For interventional studies, dietary intervention, randomization procedure, blinding of measurements, compliance with the interventions, and baseline and post-intervention gut microbiota composition were added to the checklist. For observational studies, the dietary assessment method was added to the checklist.

The full text of the papers was checked to retrieve the relevant information. The primary outcome to be investigated in this review was the effect of the Mediterranean diet on gut microbiota composition (bacterial abundance and diversity) and microbiota-derived metabolites. Secondary outcomes were the effects of MD on clinical outcomes, including prevention and treatment of weight gain and obesity, hyperglycemia, insulin resistance, inflammation, and dyslipidemia.

Quality assessment of studies

Risk of bias assessment was accomplished by one author (A. Kh.); afterward, an accuracy check was performed by another author (H-S. E.). Interventional studies' quality was assessed using the Cochrane risk of bias tool [19]. Cochrane risk of bias tool consisted of six domains: random sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective reporting, and other biases. Each interventional study was categorized as high, medium, and low risk.

The quality of observational studies was assessed using an adapted New-Castle Ottawa Scale (NOS) tool for cross-sectional and cohort studies which was developed to assess the quality of non-randomized studies [20, 21]. The adopted versions of NOS consist of three bias-evaluating sections: Selection, Comparability, and Outcome. Each section consists of further subsections, differing in two adopted NOS versions. High-quality articles were defined as \(\ge 7\) stars, medium (4-6 stars), and low (0-3 stars).

Results

Overview

A total of 1637 articles were obtained during the initial search (PubMed, Scopus, Web of Science, hand searching), of which 464 were deleted due to duplication, 1136 records did not meet the inclusion criteria or were inappropriate due to indirect relevance, or missing outcome data were also removed. Ultimately, 37 articles (17 observational and 20 interventional studies) successfully met the search criteria and were included in this systematic review (Fig. 1). The findings of these articles are summarized in Tables 1 and 2.

Fig. 1
figure 1

The Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) for included articles in the current study

Table 1 Characteristics of the included observational studies
Table 2 Characteristics of included interventional studies

Study characteristics

Observational studies

The findings of competent observational studies are demonstrated in Table 1. Of 17 included observational studies, 6 were cohort studies, and 11 were cross-sectional. The total number of patients in 16 observational studies was 7838. In 14 studies, the effects of MD on healthy patients were evaluated [1, 11, 15, 17, 22,23,24,25,26,27,28,29,30,31]. In a study by Cox et al., the impact of MD on cirrhotic patients alongside healthy patients was assessed [14]. Moreover, one study was conducted on senior patients with a high prevalence of cardiovascular diseases [32]. At last, a recently published study by Calabrese et al. was conducted on patients with nonalcoholic fatty liver disease (NAFLD) [33]. Studies were conducted in Italy, Spain, Greece, the USA, Egypt, the UK, Turkey, and Canada. The USA was the most frequent country with five articles [14, 15, 23, 30, 32]. The mean age of patients in observational studies was 33.97 years. Of 7640 cases with available gender distribution, 39.62% were male. In 15 observational studies, 16srRNA/DNA sequencing was performed to determine gut microbiota composition [1, 11, 14, 15, 17, 22,23,24,25,26, 28,29,30, 32, 33]. Two studies did not specify the microbiota assessment method [27, 31]. The main findings of observational studies are summarized in Fig. 2.

Fig. 2
figure 2

One main finding of each observational study included in the current study

Interventional studies

The findings of eligible interventional studies are summarized in Table 2. Of 20 interventional studies, three on metabolic syndrome [34,35,36], two were conducted on patients with Coronary Heart Disease [37, 38], one on type-2 diabetes mellitus [39], one on NAFLD or nonalcoholic steatohepatitis (NASH) cases [40], one on healthy patients with increased risk of developing colon cancer by definition [41], one on healthy individuals with risk factors for cardiovascular diseases [42], and other eleven studies were conducted on healthy patients [43,44,45,46,47,48,49,50,51,52,53]. Of eleven studies on healthy patients, five were conducted on patients with normal BMI [47, 49, 51,52,53], three on obese/overweight patients [44, 46, 50], and three other studies were conducted on both normal and elevated BMI patients in two different groups [43, 45, 48]. The total number of patients in interventional studies was 2402.

Of 20 interventional studies, nine were conducted in Spain [34,35,36,37,38, 40, 44, 50, 52], three in the USA [41, 49, 51], three in Italy [45, 46, 48], one in France [43], one in Australia [42], one in Portugal [39], one in Israel [53], and one was conducted on elders from five different centers in the world [47]. Of 20 interventional studies, one study was designed as triple arms [43], 11 studies as double arms [34, 35, 37, 38, 41, 44,45,46, 50, 52, 53], and seven studies as single arms [36, 39, 40, 47,48,49, 51]. Furthermore, five studies were designed as crossed-over trials [34, 35, 42, 46, 52]. Of 20 interventional articles, the 16S rRNA/DNA sequencing was carried out to identify gut microbiota composition. In the other four articles, either the method was not specifically mentioned or other methods were employed [43, 45, 52, 53]. The main findings of interventional studies are summarized in Fig. 3.

Fig. 3
figure 3

One main finding of each interventional study included in the current study

Quality assessment and risk of bias

Of 17 observational studies' quality assessed by the adapted New-Castel Ottawa scale, five studies had a moderate quality score (4-6) [17, 25, 26, 28, 32]. Twelve other studies had a high-quality score (\(\ge\) 7) [1, 11, 14, 15, 22,23,24, 27, 29,30,31, 33]. Data on observational studies' quality assessments are summarized in Tables 3 and 4.

Table 3 Adapted Newcastle-Ottawa assessment scale for cross-sectional studies
Table 4 Adapted Newcastle-Ottawa assessment scale for cohort studies

We used the term Not Applicable in different sections of the interventional studies quality assessment tool if 1) the patients were informed of their allocation to their groups (†) or 2) participants' randomized allocation was performed, but further blinding was not applicable due to the nature of dietary interventions (‡) or 3) only one group was assessed before and after dietary intervention (single arm) (). Data on interventional studies' quality assessment are summarized in Table 5.

Table 5 Cochrane bias assessment tool for interventional studies

Mediterranean Diet and gut microbiota diversity

Observational studies

Of 17 observational studies, 10 reported a correlation between MD and alpha or beta diversity [11, 14, 15, 23,24,25,26, 29, 30, 32]. Alpha diversity explains the structure of bacterial richness (number of taxonomic groups) or evenness (distribution of the abundance of groups) of both [54], while beta diversity summarizes the degree to which bacteria differ from one another [55]. In other words, the alpha index evaluates intra-sample diversity, whereas the beta index assesses inter-sample diversity.

Of 10 studies with reported diversity, eight analyzed both alpha and beta diversity [11, 15, 24,25,26, 29, 30, 32], one study reported only alpha diversity [14], and one study reported only beta diversity [23]. Four (of nine) studies reported beta diversity via the Bray-Curtis measure [23, 26, 29, 30], four reported beta diversity via the UniFrac measure [11, 15, 24, 32], and one study did not specify the beta measure [25]. Five (of nine) studies reported alpha diversity via the Shannon index [14, 15, 26, 29, 30]; one study reported alpha through four measures: Chao1, OTUs, Simpson, and Shannon [24]; one study reported Shannon and Chao1, simultaneously [25]. Another study reported alpha as both Shannon and Faith PD measures [32], and another did not specify the method for alpha diversity measurement [11].

Four (of nine) studies with a report on beta diversity had a significant bacterial separation following MD adherence [15, 23, 24, 26], three did not show a significant correlation [11, 29, 30], one had a significant correlation with some of the Mediterranean dietary components, but did not mention the correlation with total MD [32]. The last study did not mention the outcome of the beta diversity assessment [25].

Four (of nine) studies with a report on alpha diversity did not yield a significant correlation with MD adherence [11, 26, 29, 30], whereas three reported that MD adherence resulted in higher bacterial diversity [14, 15, 24]. In one study, the correlation between MD adherence and alpha diversity was significant via the Shannon index yet insignificant through Faith's PD [32]; another study revealed a significant correlation via Chao1, yet insignificant via Shannon [25]. Microbiota diversity in observational studies is summarized in Table 1.

Interventional studies

Of 20 interventional studies, 14 investigated alpha and beta diversity [34, 36, 38,39,40,41,42, 44, 46, 48,49,50,51,52], and two study only assessed alpha diversity [47, 53]. Of 16 studies with a report on alpha diversity, 13 claimed no significant association between alpha diversity and MD adherence [34, 38,39,40,41,42, 44, 46,47,48,49,50, 52]. Only in three studies, there were a significant association between MD and alpha diversity [36, 51, 53]. Furthermore, of 14 articles with a report on beta diversity, nine did not report any significant separation utilizing beta diversity neither between case and control group nor within a group before and after intervention [34, 38, 39, 42, 44, 46, 49,50,51], three study reported a significant bacterial separation after MD intervention [36, 40, 41]. Another study reported a significant difference at baseline between the case and control group before any intervention occurred [48]. Finally, in one study, MD adherence significantly affected beta diversity in intervened cases compared to the control group [52]. Microbiota diversity in interventional studies is summarized in Table 2.

Mediterranean Diet and different bacterial abundance

Observational studies

Of 17 observational studies, 16 reported significant effects of MD on microbiota composition, and the abundance of at least one bacterium at the phylum, genus, or species level differed between groups. Just one study did not have a significant finding (p-value>0.05), but still, they claimed that there was a trend toward increasing Firmicutes and decreasing Bacteroidetes with MD adherence [28]. However, in that study, the specific impact of physical activities on microbiota, instead of MD adherence, is delineated and the impact is even statistically significant (p<0.05). Of 16 studies with a significant report on microbiota abundance, an increase in Faecalibacterium genus was reported in four articles [22, 27, 30, 31]. Furthermore, four articles reported an increase in either Bacteroidetes phylum or Bacteroides genus [1, 17, 26, 30]. Four articles mentioned an increase in either Prevotellacea family or Prevotella genus [11, 17, 23, 32]. The results of observational studies regarding gut bacterial abundance are summarized in Table 1.

Interventional studies

Of 20 interventional studies, 17 reported a significant change in bacterial abundance after MD intervention in at least one bacterium at the phylum, genus, or species level. Only three studies failed to find a significant change in bacterial abundance [35, 41, 43]. Prevotella, either in genus or species level, was increased in four studies [38, 39, 47, 48] yet decreased in one study [44]. Nevertheless, in one of the four studies with a report on increase in Prevotella, the amount of increase was not statistically significant [39]. In four studies, Faecalibacterium was increased at the genus level [37, 38, 40, 47]. Firmiticus phylum was increased in three studies [36, 39, 51], whereas it decreased in one [48]. In one study, both increasing and decreasing trends were observed in members belonging to Firmicutes Phylum [50]. Results of interventional studies regarding gut bacterial abundance are summarized in Table 2.

Effect of Mediterranean diet adherence on microbial metabolites

Observational studies

Of 17 observational studies, 11 reported a significant change in microbial metabolites in MD adherent participants [1, 11, 14, 17, 22, 23, 25, 27, 29, 32, 33]. Five articles reported a significant increase in main SCFAs following MD adherence [11, 23, 25, 27, 29]. Acetate was significantly increased in four studies [1, 11, 23, 25], while in one study, it was significantly increased via the MDS assessment tool yet decreased via HEI [32]. Propionate was increased remarkably in five studies [11, 17, 23, 25, 32]. Microbial metabolites in observational studies are summarized in Table 1.

Interventional studies

Nine of 20 interventional studies showed significant changes in microbiota-derived metabolites following Mediterranean dietary intervention [34,35,36, 44, 46, 47, 49, 52, 53]. One article mentioned a remarkable increase in the concentration of SCFAs [47], and two reported a significant increase in propionic acid following MD [46, 49]. Microbial metabolites in interventional studies are summarized in Table 2.

Effect of Mediterranean diet adherence on clinical outcomes

Observational studies

Four studies reported a significant clinical or clinical-related laboratory outcome [1, 14, 27, 33]. An increase in fecal moisture and defecation frequency, a decrease in bloating [1], decreasing a 90-day hospitalization risk [14], better glycemic/hyperlipidemic state control [33], and decreasing serum IL-8 level [27] were clinical outcomes mentioned in observational studies.

Interventional studies

In total, 15 studies reported a significant clinical outcome following dietary intervention [34,35,36, 39, 40, 42,43,44,45,46,47,48, 50, 52, 53]. Lowering inflammation was reported in four articles [43, 45,46,47]. Optimized diabetic control was reported eight times [34,35,36, 39, 40, 44, 50, 53]. Decreasing fat mass was reported four times [36, 48, 50, 53], lowering systolic blood pressure reported once [42], and better bowel movement was reported once [52].

Discussion

This systematic review aimed to summarize the results of observational and interventional studies that examined the efficacy of the MD on the gut microbiota composition and clinical outcomes in different groups of people with distinct demographic characteristics and health statuses. This study reviewed 37 documents, divided into interventional and observational studies.

Consumption of the Mediterranean diet is associated with a different microbiota composition compared to Western-type dietary patterns. The microbiota composition associated with MD is characterized by higher microbial biodiversity. This characteristic of gut microbiota is defined as "α-diversity," demonstrating the number of species present in the microbiota and is associated with the health of individuals [54]. Besides, an intersample bacterial separation between two groups is measured through beta diversity [55]. In a study by Bowyer et al., Alpha diversity was significantly increased following MD adherence [24]. In another study by Maskarinec et al., alpha diversity was assessed in four dietary indices: HEI-2010, aHEI-2010, aMed, and DASH. Alpha diversity was increased significantly in tertiles in all four dietary indices [15].

Animal and human studies on gut microbiota composition via fecal samples have shown that all dietary changes could modulate gut microbial composition. In healthy subjects, a balanced diet can induce the formation of good microbial flora, which consists of all species of bacteria living in a system of control and mutual balance [56].

It is well established that gut microbial alteration may affect metabolism via secreted metabolites. The fermentation of the dietary components of the MD by the gut bacteria leads to the production of specific metabolites, such as SCFA, which is represented in the feces of subjects that follow MD [57]. SCFAs are carboxylic acids with six carbon atoms, maximumly, more frequently including acetic, propionic, and butyric acids [58].

Although the role of genetics in obesity is well known to everyone, human microbiota also plays a crucial role [59]. SCFA level, as the main metabolites of gut microbiota, is known to be altered in obesity as a result of dysbiosis, with more abundant Firmicutes relatively [60, 61]. Furthermore, SCFA alteration in obese patients results from increased Lactobacillus and Staphylococcus [62] and decreased Bifidobacterium [63].

The highest colon-rectal levels of SCFAs, specifically butyrate, would contribute to the reduced risk of CRC observed in Mediterranean countries. These protective effects could also contribute to the reduced presence of Fusobacterium nucleatum, which is mainly present in the colon of patients with CRC, and based on some related studies, it could be associated with the onset of this cancer [64]. Seven articles in our study reported increased SCFAs following MD (Tables 1 and 2).

On the other hand, TMAO (trimethylamine N-oxide) metabolites are present in higher concentrations in subjects that follow a Western diet [65]. Surprisingly, in a study by Barber et al., TMAO increased 1.5 times after MD. They speculated that ingesting choline-riched plant food, including legumes, prior to urinary sampling in the MD group might have been a reasonable explanation [52].

Faecalibacterium prausnitzii, a main known butyrate-producing bacteria with anti-inflammatory effects [66], was increased in seven studies [22, 27, 30, 37, 38, 47, 53], and no decrease in its level was reported in any of the included studies, despite the debate on impacts of MD on F. Prausnitzii level in the previous document [55]. Our findings were also parallel to those claiming MD may increase bacteria with polysaccharide affinity, including E. Eligens [30, 47], Roseburia species [29, 38, 44, 47], Butyricicoccus species [29, 49, 52] and also may decrease bacteria with simple sugars affinity [55], including C. aerofaciens [47]. Eventually, Bacteroides [1, 17, 26, 30, 37, 44, 47, 48] and Parabacteroides [37, 44] were among the most frequent microbiota species, which increased following MD adherence.

A significant effect on gut microbiota composition is believed to require long-term dietary pattern intervention. A study by Djuric et al. was conducted on the mucosal bacterial flora of the colon before or after six months of the Mediterranean or Western-type experimental diet. It revealed no significant differences in microbiota before or after the intervention [41]. Hence, a consistent dietary intervention should be considered for an almost permanent beneficial microbiota alteration.

In a study conducted on patients with metabolic syndrome, consuming the Mediterranean or traditional diet for two years, the MD has shown that it could partially reduce the typical dysbiosis of metabolic syndrome. The authors observed an increase in Bifidobacterium genera of the MD group [37].

Another outcome investigated in this review was the effect of the MD on the related clinical outcomes. In total, 19 studies investigated clinical outcomes after MD. The main reported outcomes were better diabetes mellitus management in nine articles, lowering inflammation in five articles, lowering fat mass in five, increasing bowel movement in one, and lowering hospitalization risk in one article (Tables 1 and 2).

Despite some controversies on the effect of SCFAs on inflammation [67], most studies in the literature delineated that SCFAs can decrease inflammation in the human body via inhibiting TNF-alpha and also upregulating IL-10 as an anti-inflammatory cytokine [68, 69]. Some authors also claimed that SCFAs can induce apoptosis, interrupt leukocyte migration, and inhibit the production of inflammatory mediators [70]. Besides, due to the epigenetic effects of SCFAs and their interaction with tissue receptors, their beneficial impacts on glucose homeostasis and decreasing glucose resistance have been proposed [71].

Strengths and limitations

In this study, we thoroughly investigated observational and interventional studies, and clinical and microbiota alteration were assessed as outcomes simultaneously. However, most of the included studies in this systematic review had a limited number of participants, and only 15 of 37 articles investigated more than 100 participants in their research, which makes them heterogeneous in their design and representativeness. Various dietary assessment methods in observational studies and dietary interventions in interventional studies were utilized. Furthermore, some authors defined MD as a monounsaturated fatty acid-rich or enriched diet, while others adjusted this diet by adding nuts or other foods. Performing meta-analysis was impossible due to the included studies' heterogeneous nature.

Conclusion

Adherence to the MD is associated with significant beneficial changes in the gut microbiota diversity, composition, and functions and major clinical improvements in most populations.