Association Between Adherence Levels to the EAT-Lancet Diet in Habitual Intake and Selected Gut Bacteria in a Mexican Subpopulation

Abstract

The EAT-Lancet diet is an outstanding model of a healthy, environmentally sustainable diet. However, its effects on the gut microbiota remain poorly explored. This study assessed the asso-ciation between adherence to the EAT-Lancet diet in habitual intake and the relative abundance of selected gut bacteria in a Mexican subpopulation. Fifty-four young adults (18–35 years) completed a validated Food Frequency Questionnaire (FFQ) and were nutritionally assessed. Participants were grouped into low, moderate, and high adherence levels to the EAT-Lancet diet. Blood samples were analysed for glucose and lipid profiles, and gDNA from faecal samples was analysed using Real-time qPCR to quantify gut bacteria. While no significant differences in bacterial abundance were observed across adherence levels, correlations emerged with increased adherence. Notably, Bifidobacterium negatively correlated with beef and lamb intake (rho −0.5, p < 0.05), and Akkermansia muciniphila negatively correlated with fish intake (rho −0.8, p < 0.05). Bilophila wadsworthia positively correlated with triglycerides, while Prevotella copri and Faecalibacterium prausnitzii negatively correlated with body fat and blood pressure, respectively. In addition, a non-significant trend toward a higher abundance of Firmicutes, Akkermansia muciniphila, and Prevotella copri was observed in the high-adherence group, whereas Lactobacillus tended to be more abundant in participants with low and moderate adherence. These findings suggest that adherence to the EAT-Lancet diet modulates gut microbiota composition. However, further controlled interventional studies are needed to confirm these effects and their implications for human health.

1. Introduction

Diet plays a key role in shaping the gut microbiota, a dynamic community of microorganisms essential for human and animal health [1,2,3]. The composition and functionality of the gut microbiota are directly influenced by the quality, quantity, and type of consumed foods, impacting metabolic, immunological, and neurological processes [3,4]. Recent studies have emphasised how specific dietary patterns can promote or alter microbial diversity, potentially influencing the development of inflammatory, metabolic, and digestive diseases [5,6].

In recent decades, the global climate and nutritional crises have posed significant challenges to food systems and public health. Overexploitation of natural resources, climate change, and the rising prevalence of non-communicable diseases have underscored the need to align traditional dietary recommendations with sustainable and health-promoting objectives [7,8]. In this context, the concept of sustainable diets has emerged, emphasising eating patterns that not only support human health but also minimise environmental impacts and respect planetary boundaries [9].

One of the most outstanding examples of a sustainable diet model is the EAT-Lancet diet, introduced in 2019 by the EAT-Lancet Commission in their report, “Food in the Anthropocene: the EAT-Lancet Commission on Healthy Diets from Sustainable Food Systems” [10]. This dietary model aims to balance human health with environmental sustainability by defining clear limits for food systems to operate within planetary boundaries. The EAT-Lancet diet emphasises a high intake of plant-based foods, such as fruits, vegetables, whole grains, nuts, and legumes, while limiting the consumption of animal-based products, added sugars, and saturated fats [10,11].

The EAT-Lancet approach is founded on two core principles: promoting human health and respecting planetary limits. From a health perspective, this diet seeks to reduce the global burden of diet-related diseases, including obesity, diabetes, and cardiovascular diseases. From an environmental perspective, it aims to minimise greenhouse gas emissions, reduce freshwater use, and limit biodiversity loss, ensuring that the global food system operates within the Earth’s ecological capacity [10,11].

While the EAT-Lancet diet has been extensively studied regarding environmental sustainability and health outcomes, its effects on the gut microbiota remain an emerging area of research [12,13]. As dietary patterns highly influence the gut microbiota, it represents a critical link between nutrition, health, and well-being. Emerging studies have started exploring the effects of adhering to this diet on the gut microbiome [12,13]. However, the gut microbiota’s response to adherence to the EAT-Lancet diet may vary based on the genetic, cultural, and dietary characteristics of specific populations, such as the Mexican population, whose traditional dietary patterns differ significantly from this model [14].

This study explored the association between the adherence levels to the EAT-Lancet diet in a real-life habitual intake context, and the relative abundance of selected gut bacteria in a Mexican subpopulation. The bacteria included the main phyla present in the gut microbiota, Firmicutes and Bacteroidetes (recently called Bacillota and Bacteroidota, respectively), representing 90% of the total gut microbiota composition [3]. Health-beneficial bacteria, such as Lactobacillus, Bifidobacterium, Akkermansia muciniphila, and Faecalibacterium prausnitzii, were also included, especially because of their capacity to produce Short-Chain Fatty Acids (SCFAs) and their established role in metabolic health [6,14,15,16]. Moreover, bacteria with conflicting effects on human health, such as Prevotella copri, Streptococcus thermophilus, and Clostridium coccoides [4,17,18,19,20], were included. Additionally, Bilophila wadsworthia was included, as it has been recognised for its association with meat consumption and pro-inflammatory effects [19,21].

Unlike previous studies that have explored the EAT-Lancet diet in controlled interventions or in populations with habitual intake but distinct cultural backgrounds [12,13], this study uniquely assesses its association with the gut microbiota in a Mexican subpopulation, whose current diet differs significantly from the planetary health model. By evaluating adherence levels within a real-life dietary context and linking them to the relative abundance of health-relevant gut bacteria, our research contributes novel evidence on the potential microbiota-related effects of sustainable dietary patterns in different populations. These findings may inform culturally tailored strategies to promote gut and planetary health.

2. Materials and Methods

2.1. Study Design and Participants

A cross-sectional study was conducted using a subsample of 54 young adults (aged 18–35 years) from the baseline data of the NutriSOS study [22]. As the participants were selected from an existing cohort, a convenience sampling strategy was applied. Recruitment for the NutriSOS study was conducted at local universities and public health centres in Ciudad Guzmán, Jalisco, Mexico, using posters, digital invitations, and word-of-mouth referrals. All the participants provided informed consent to participate in this research. The inclusion and exclusion criteria required that the participants had no diagnosed medical conditions (e.g., metabolic, immunological, gastrointestinal), were not undergoing pharmacological treatment, and had not consumed antibiotics in the three months prior to the assessment. Detailed inclusion and exclusion criteria are provided in Supplementary Material S1 (Table SM1.1).

2.2. Socioeconomic and Demographical Characteristics of the Sample

For the socioeconomic data, the National Autonomous University of Mexico educational level classification and the occupation level classification of INEGI [23] were grouped into high, medium, and low [24]. The Mexican Association of Market Research Agencies was used to classify monthly income [25].

2.3. Anthropometric and Body Composition Evaluation

Body weight and body composition were assessed using specialised bioelectrical impedance equipment (TANITA BC-601 FITSCAN, Arlington Heights, IL, USA). Height was measured with a Smartmet^® stadiometer following the Frankfurt plane, with the participants barefoot. Waist and hip circumferences were evaluated using a Lufkin^® metal tape measure: waist circumference was measured at the midpoint between the lower rib and the iliac crest at the end of normal expiration, and hip circumference was measured at the widest point [26]. Neck circumference was assessed following the International Society for the Advancement of Kinanthropometry (ISAK) protocol [27].

All measurements were performed by a certified ISAK Level 2 nutritionist, adhering to the protocol described by Suverza and Haua [26]. The participants were instructed to remove their shoes, wear light clothing, and maintain a minimum fasting period of two hours. The women were advised to avoid measurements during their menstruation to prevent hydration-related alterations that could affect bioimpedance results. Additionally, all metallic objects were removed prior to the assessments.

2.4. Dietary Intake and Adherence to the EAT-Lancet Diet

A validated and adapted 245-item Food Frequency Questionnaire (FFQ) was administered using Nutriecology^® software [28]. The foods consumed were categorised into 35 food groups based on the EAT-Lancet dietary classification (22 food groups) and additional commonly consumed foods in Mexico (13 food groups) [10]. Details of the food items and their group classifications are provided in Supplementary Material S2 (Table SM2.1). To minimise errors in estimating portion sizes and intake frequencies, trained nutritionists conducted the questionnaires using food replicas, portion images, measuring cups, and scoops. Caloric and nutrient intakes were automatically calculated in Nutriecology^®, which utilises Mexican food composition tables [28,29,30].

Adherence to the EAT-Lancet diet was determined by adjusting dietary intake to 2500 kcal per person per day and calculating the EAT-Lancet Dietary Index [31]. A maximum score of 18 could be obtained, and adherence levels were classified into tertiles: low (7 to 9 points), moderate (10 to 11 points), and high (12 to 13 points). For foods assessed in their cooked state (e.g., rice, maize, wheat, beans, lentils, chickpeas), correction factors were applied to convert them to their raw equivalent [32]. Dietary intake references are detailed in Supplementary Material S3 (Table SM3.1).

2.5. Physical Activity Assessment and Classification

Physical activity was assessed using the International Physical Activity Questionnaire (IPAQ), which collected information on the type, frequency, duration, and intensity of activities performed. The type of physical activity was divided into three categories [33]:

• Mild aerobic activities: Activities such as walking without elevating the heart rate beyond 50% of maximum capacity.
• Moderate to intense aerobic activities: Activities such as brisk walking, jogging, running, swimming, aerobics, and cycling, which elevate the heart rate to 50–90% of maximum capacity.
• Anaerobic activities: Weightlifting sessions lasting at least 20 min per day.

The level of physical activity was classified according to IPAQ scoring criteria [34]:

• Low activity: Participants who did not meet the criteria for moderate or high levels of physical activity.
• Moderate activity: Participants who met at least one of the following criteria:
  • Vigorous-intensity activity on 3 or more days for at least 20 min per day.
  • Moderate-intensity activity or walking on 5 or more days for at least 30 min per day.
  • A combination of walking, moderate-intensity, or vigorous-intensity activities on 5 or more days, achieving a minimum of 600 MET-minutes/week.
• High activity: Participants who met one of the following criteria:
  • Vigorous-intensity activity on at least 3 days, accumulating at least 1500 MET-minutes/week.
  • A combination of walking, moderate-intensity, or vigorous-intensity activities on seven or more days, achieving a minimum of 3000 MET-minutes/week.

2.6. Clinical and Metabolic Biomarkers

Systolic and diastolic blood pressure were measured using a Medstar^® sphygmomanometer in accordance with Mexican regulations [35]. For the determination of metabolic biomarkers, blood samples were collected following an 8–12 h overnight fast and centrifuged at 3500 rpm for 15 min to separate the serum. Aliquots were prepared and stored at −80 °C until analysis. Fasting glucose (Cat. No. 1001190), triglycerides (Cat. No. 1001313), total cholesterol (Cat. No. 41022), low-density lipoprotein (LDL) cholesterol (Cat. No. BSIS51-E), and high-density lipoprotein (HDL) cholesterol (Cat. No. BSIS37-E) were quantified using enzymatic colorimetric methods with commercial kits from Spinreact^® S.A/S.A.U (Girona, Spain) [36].

2.7. Bacteria Identification

2.7.1. Stool Collection

Each participant received a stool collection kit consisting of a wide-mouth sterile container, a glove, and a tongue depressor packaged in a resealable plastic bag along with a digital guide (Supplementary Material S4). Stool samples were aliquoted into 1.5 mL Eppendorf tubes and immediately stored at −80 °C. A detailed description of the collection method is provided in Supplementary Material S5.

2.7.2. DNA Extraction from Faeces

Bacterial DNA extraction and purification were performed using the QIAamp DNA Stool Mini Kit (51604, QIAGEN, Hilden, Germany) for rapid genomic DNA isolation. The process involves two main stages: lysis and impurity separation, followed by DNA purification using spin columns. The concentration and purity of the extracted DNA from faeces were assessed via the 260/280 nm absorbance ratio using a microplate spectrophotometer (Multiskan Sky, Thermo Fisher Scientific, Waltham, MO, USA). The manufacturer’s protocol was followed with some modifications, as detailed in Supplementary Material S6.

2.7.3. Identification of Gut Bacteria

Gut bacteria were identified using Real-time qPCR on the QuantStudio 5 platform (Applied Biosystems, Waltham, MO, USA). The selected bacteria included Firmicutes and Bacteroidetes phyla; Lactobacillus and Bifidobacterium genera; and the species Akkermansia muciniphila, Faecalibacterium prausnitzii, Prevotella copri, Bilophila wadsworthia, Streptococcus thermophilus, and Clostridium coccoides. Specific primers previously validated in the literature (Table 1) were synthesised by the company Integrated DNA Technologies (IDT, Coralville, IA, USA). Primer specificity and target amplification were confirmed via in silico analysis using the Primer3 tool on the NCBI Primer-BLAST platform [37]. Primers were hydrated with sterile water and pipetted at a volume 10 times their concentration in nanomoles, as per the manufacturer’s instructions [38,39].

The detection chemistry used for the qPCR was the Fast SYBR™ Green Master Mix (Applied Biosystems by Thermo Fisher Scientific, Vilnius, Lithuania), containing the intercalating agent SYBR^® Green I Dye, the enzyme AmpliTaq^® Fast DNA Polymerase UP (Ultra-Pure), the enzyme Uracil-DNA Glycosylase (UDG), the passive reference ROX™, dNTPs, and an optimised reaction buffer.

Reaction concentrations were prepared following the manufacturer’s guidelines with adjustments from preliminary optimisation tests. Each reaction included 5 μL of master mix, 0.2 μL of each primer at a 20 µM concentration (forward and reverse), and a variable volume of DNA adjusted to achieve 50 ng per qPCR reaction. The total reaction volume was standardised to 10 μL using molecular biology-grade water.

Pilot tests were conducted to optimise thermocycling conditions based on previously published protocols (Table 1) and the studies by Rodríguez-Lara et al. [18], Salaiza Ambriz [38], and Ramírez Guerrero [39]. For sequences with shorter amplicons (<1000 bp), the cycling conditions included initial denaturation at 95 °C for 20 s, annealing at 95 °C for 3 s, and elongation at 60 °C for 30 s for a total of 40 cycles. For sequences with longer amplicons (>1000 bp), such as Bilophila wadsworthia, an additional elongation step at 72 °C for 40 s was added after 40 cycles of 95 °C for 3 s and 60 °C for 30 s [21,39]. Melting curve analysis followed conditions of 95 °C for 10 s, 60 °C for 1 min, and 95 °C for 10 s.

To standardise the method, an initial run was performed for each bacterium to verify amplification under the established conditions. Melting curves were analysed, and efficiency curves were generated using 10-fold dilutions for all primer sets. The results confirmed that the standardisation met optimal conditions [38,39]. All samples were analysed in duplicate, and non-template controls were used to assess run reliability in each qPCR. If duplicates deviated by more than 2 cycles, the reaction was repeated [18].

2.7.4. Relative Abundance Quantification

The relative abundance of each bacterium was calculated using a universal primer (Table 1) as a reference, applying the following formula [40,41]:

$$2^{-∆Ct}{=2}^{-[Ct\left(Target Bacteria\right)-Ct(Universal Primer\left)\right]}$$

(1)

where

• The factor of 2 reflects the theoretical assumption that each PCR cycle results in an exponential doubling of the DNA present in the original sample.
• ∆Ct = The delta symbol (∆) represents the difference between the Ct values of the specific bacterium and the universal reference primer.

Table 1. Primer pairs used for each bacterium in Real-time qPCR.

Bacteria	Primer	Target Gene	Number of Bases	Sequence	Amplicon Size	Reference
Firmicutes (Bacillota)	Forward	16S rRNA	21	5′-TGAAACTCAAAGGAATTGACG-3	200	[39,42]
	Reverse		17	5′-ACCATGCACCACCTGTC-3′
Bacteroidota (Bacteroidetes)	Forward	16S rRNA	20	5′-CAAACAGGATTAGATACCCT-3′	240	[39,42]
	Reverse		19	5′-GGTAAGGTTCCTCGCGTAT-3′
Lactobacillus	Forward	16S rRNA	19	5′-AGCAGTAGGGAATCTTCCA-3′	341	[14,39]
	Reverse		17	5′-CACCGCTACACATGGAG-3′
Bifidobacterium	Forward	16S rRNA	18	5′-TCGCGTCCGGTGTGAAAG-3′	243	[14,38]
	Reverse		17	5′-CCACATCCAGCATCCAC-3′
Akkermansia muciniphila	Forward	16S rRNA	20	5′-CAGCACGTGAAGGTGGGGAC-3′	329	[38,43]
	Reverse		20	5′-CCTTGCGGTTGGCTTCAGAT-3′
Faecalibacterium prausnitzii	Forward	16S rRNA	19	5′-GGAGGAAGAAGGTCTTCGG-3′	248	[16,39]
	Reverse		21	5′-AATTCCGCCTACCTCTGCACT-3′
Prevotella copri	Forward	16S rRNA	20	5′-CCGGACTCCTGCCCCTGCAA-3′	106	[17,38]
	Reverse		20	5′-GTTGCGCCAGGCACTGCGAT-3′
Clostridium Coccoides	Forward	16S rRNA	19	5′-CGGTACCTGACTAAGAAGC-3′	429	[18,38]
	Reverse		19	5′-AGTTTCATTCTTGCGAACG-3′
Streptococcus thermophilus	Forward	16S rRNA	22	5′-TTATTTGAAAGGGGCAATTGCT-3′	281	[20,38]
	Reverse		21	5′-GTGAACTTTCCACTCTCACAC-3′
Bilophila wadsworthia	Forward	tpa	26	5′-CAACGTCCCCACCATCAAGTTCTCTG-3′	1001	[21,39]
	Reverse		26	5′-TGAATTCGCGGAAGGAGCGAGAGGTC-3′
Universal	Forward	16S rRNA	20	5′-AAACTCAAAGGAATTGACGG-3′	180	[42,44]
	Reverse		18	5′-CTCACAACACGAGCTGAC-3′

Table 1. Primer pairs used for each bacterium in Real-time qPCR.

Bacteria	Primer	Target Gene	Number of Bases	Sequence	Amplicon Size	Reference
Firmicutes (Bacillota)	Forward	16S rRNA	21	5′-TGAAACTCAAAGGAATTGACG-3	200	[39,42]
	Reverse		17	5′-ACCATGCACCACCTGTC-3′
Bacteroidota (Bacteroidetes)	Forward	16S rRNA	20	5′-CAAACAGGATTAGATACCCT-3′	240	[39,42]
	Reverse		19	5′-GGTAAGGTTCCTCGCGTAT-3′
Lactobacillus	Forward	16S rRNA	19	5′-AGCAGTAGGGAATCTTCCA-3′	341	[14,39]
	Reverse		17	5′-CACCGCTACACATGGAG-3′
Bifidobacterium	Forward	16S rRNA	18	5′-TCGCGTCCGGTGTGAAAG-3′	243	[14,38]
	Reverse		17	5′-CCACATCCAGCATCCAC-3′
Akkermansia muciniphila	Forward	16S rRNA	20	5′-CAGCACGTGAAGGTGGGGAC-3′	329	[38,43]
	Reverse		20	5′-CCTTGCGGTTGGCTTCAGAT-3′
Faecalibacterium prausnitzii	Forward	16S rRNA	19	5′-GGAGGAAGAAGGTCTTCGG-3′	248	[16,39]
	Reverse		21	5′-AATTCCGCCTACCTCTGCACT-3′
Prevotella copri	Forward	16S rRNA	20	5′-CCGGACTCCTGCCCCTGCAA-3′	106	[17,38]
	Reverse		20	5′-GTTGCGCCAGGCACTGCGAT-3′
Clostridium Coccoides	Forward	16S rRNA	19	5′-CGGTACCTGACTAAGAAGC-3′	429	[18,38]
	Reverse		19	5′-AGTTTCATTCTTGCGAACG-3′
Streptococcus thermophilus	Forward	16S rRNA	22	5′-TTATTTGAAAGGGGCAATTGCT-3′	281	[20,38]
	Reverse		21	5′-GTGAACTTTCCACTCTCACAC-3′
Bilophila wadsworthia	Forward	tpa	26	5′-CAACGTCCCCACCATCAAGTTCTCTG-3′	1001	[21,39]
	Reverse		26	5′-TGAATTCGCGGAAGGAGCGAGAGGTC-3′
Universal	Forward	16S rRNA	20	5′-AAACTCAAAGGAATTGACGG-3′	180	[42,44]
	Reverse		18	5′-CTCACAACACGAGCTGAC-3′

2.8. Statistical Analysis

Data normality was assessed using the Shapiro–Wilk test. Descriptive analysis was conducted, with the results reported as means and standard deviations. For comparisons according to diet adherence level, the Kruskal–Wallis test with Dunn’s post hoc analysis was applied to non-normally distributed variables, while One-way ANOVA with Bonferroni post hoc analysis was used for normally distributed data. Categorical variables were analysed using the Chi-squared test. Spearman correlations were conducted to evaluate correlations among variables. To control for multiple comparisons, we applied the Bonferroni correction, which provides a stringent control of Type I errors. Although conservative, this approach ensures that only the most robust correlations remain statistically significant [45]. All statistical analyses were performed using STATA v12, and graphs were generated in GraphPad Prism 10.

2.9. Ethics Considerations

This study was approved by the Ethics Committee of the Centre for Studies and Research in Behaviour at the University Centre for Biological and Agricultural Sciences (CUCBA), University of Guadalajara (reference: CUCBA/CEIC/CE/002/2022), and by the Technical Research Committee of the University Centre of the South (CUSUR) (reference: 2021D001). The protocol from which samples were obtained, was registered on ClinicalTrials.gov (NCT05457439). This study adhered to the guidelines of the Declaration of Helsinki. All the participants were adults who provided informed consent before being included in this study.

3. Results

3.1. Socioeconomic Characteristics of the Sample According to Adherence Level

The sociodemographic characteristics of the sample, according to adherence level to the EAT-Lancet diet, are summarised in

Table 2. Sociodemographic and socioeconomic characteristics of the sample according to adherence level to the EAT-Lancet diet.

Variable	General Subsample	Adherence Level
		Low	Moderate	High	p-Value
		n (%)	n (%)	n (%)
	n = 54	20 (37.00) ^	20 (37.00) ^	14 (26.00) ^
Sex	n (%)
Women	38 (70.37)	12 (60.00)	13 (65.00)	13 (92.86)	0.095
Men	16 (29.63)	8 (40.00)	7 (35.00)	1 (7.14)
Educational level
Basic	0 (0.00)	0 (0.00)	0 (0.00)	0 (0.00)	0.865
Medium	34 (62.96)	14 (70.00)	12 (60.00)	8 (57.14)
Higher	20 (37.04)	6 (30.00)	8 (40.00)	6 (42.86)
Occupational level
Low	22 (40.74)	10 (50.00)	7 (40.00)	5 (35.71)	0.626
Medium	8 (14.81)	6 (30.00)	1 (5.00)	1 (7.14)
High	24 (44.44)	4 (20.00)	12 (55.00)	8 (54.14)
Monthly income
0–2699	12 (22.22)	5 (25.00)	3 (15.00)	4 (28.57)	0.300
2700–6799	19 (35.19)	9 (45.00)	6 (30.00)	4 (28.57)
6800–11,599	10 (18.52)	4 (20.00)	2 (10.00)	4 (28.57)
11,600–34,999	12 (22.22)	2 (10.00)	8 (40.00)	2 (14.29)
35,000–84,999	1 (1.85)	0 (0.00)	1 (5.00)	0 (0.00)
+85,000	0 (0.00)	0 (0.00)	0 (0.00)	0 (0.00)
Age
Average age	24.70	23.95	25.20	25.07	0.340
Standard deviation	4.29	4.22	3.80	5.136
Minimum	19.00	20	19	19
Maximum	35.00	35	30	34

Note: ^% from the total sample; p < 0.05 was considered statistically significant and was obtained from the Chi-squared test.

. The majority of the participants were women (70%), and most belonged to medium to high socioeconomic levels. No significant differences were observed among the adherence groups regarding socioeconomic variables (p > 0.05). However, the high-adherence group showed a trend toward higher education and occupational level.

3.2. Anthropometric, Clinical, Biochemical, and Body Composition Data of the Sample According to Their Adherence Level to the EAT-Lancet Diet

As shown in

Table 3. Anthropometric, clinical, biochemical, and body composition data of the sample according to adherence level to the EAT-Lancet diet.

Variable	General Subsample		Adherence Level						p-Value
			Low		Moderate		High
			n (%)		n (%)		n (%)
	n = 54		20 (37.00) ^		20 (37.00) ^		14 (26.00) ^
	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Anthropometric data
Height (cm)	164.24	8.00	164.59 a	7.79	164.80 a	8.21	162.96 a	8.46	0.7888
Weight (kg)	71.10	20.08	75.40 a	25.80	70.05 a	17.03	66.46 a	13.89	0.6916′
BMI (kg m−2)	26.07	5.89	27.35 a	7.37	25.62 a	5.05	24.91 a	4.50	0.7277′
Waist circumference (cm)	82.09	13.98	84.12 a	17.04	82.25 a	12.92	78.98 a	10.55	0.7207′
Hips circumference (cm)	100.03	16.87	104.28 a	15.01	95.33 a	21.63	100.69 a	9.24	0.6186′
Waist–hip ratio	0.93	0.90	0.80 a	0.07	1.15 a	1.47	0.78 a	0.06	0.1956′
Neck (cm)	35.07	4.14	36.07	4.85	35.50	3.96	33.04	2.53	0.0923
Body composition data
Body fat (%)	30.85	9.00	31.06 a	11.16	29.66 a	6.98	32.27 a	8.53	0.4963
Visceral fat (kg)	5.02	4.20	6.15 a	5.77	4.80 a	3.14	3.71 a	2.23	0.4766′
Muscle mass (kg)	45.86	10.86	47.94 a	12.49	46.43 a	10.96	42.08 a	7.39	0.3039′
Water (%)	50.63	5.91	50.48	7.32	51.19	4.53	50.06	5.77	0.5852
Metabolic rate (kcal)	2344.89	531.76	2440.05 a	620.31	2371.50 a	541.13	2170.93 a	338.90	0.3250′
Metabolic age	40.07	21.82	43.95 a	26.89	38.05 a	18.74	37.43 a	18.25	0.8472′
Biochemical data
Glucose (mg/dL)	89.54	8.34	90.85 a	7.71	88.75 a	8.00	88.79 a	9.93	0.6825
Triglycerides (mg/dL)	94.17	51.49	98.85 a	58.15	97.50 a	51.83	82.71 a	41.84	0.4579′
Total cholesterol (mg/dL)	156.11	30.60	160.82 a	36.68	153.32 a	32.38	153.39 a	16.32	0.6954
LDL cholesterol (mg/dL)	105.23	28.77	110.33 a	31.10	103.65 a	33.25	100.22 a	16.54	0.6203
HDL cholesterol (mg/dL)	50.88	10.34	50.49 a	11.15	49.68 a	9.63	53.17 a	10.52	0.5820
Clinical data
Systolic blood pressure (mg/Hm)	104.35	15.24	108.00 a	13.61	106.25 a	17.84	96.43 a	10.82	0.0702
Diastolic blood pressure (mg/Hm)	72.22	11.06	75.75 a	10.67	72.25 a	13.03	67.14 a	6.11	0.0804
Physical activity
Days per week	2.54	2.13	3.05 a	1.96	2.15 a	2.39	2.36 a	1.95	0.3875
Minutes per day	45.19	44.67	49.00 a	41.79	44.50 a	51.96	40.71 a	39.90	0.7419
Physical activity level	n	%	n	%	n	%	n	%
Low	25.00	46.29	8.00	40.00	11.00	55.00	6.00	42.85	0.457 º
Medium	25.00	46.29	10.00	50.00	8.00	40.00	7.00	50.00
Intense	4.00	7.40	2.00	10.00	1.00	5.00	1.00	7.14
Physical activity type
Mild aerobic	21.00	38.88	6.00	30.00	11.00	55.00	4.00	28.57	0.751 º
Moderate to intense aerobic	22.00	40.74	10.00	50.00	4.00	20.00	8.00	57.12
Anaerobic	11.00	20.37	4.00	20.00	5.00	25.00	2.00	14.28

Note: ^% from the total sample. SD = standard deviation. p < 0.05 is considered statistically significant. p values were obtained from the Kruskal–Wallis test with Dunn’s post hoc analysis for non-normal data and from ANOVA with Bonferroni post hoc analysis for normal data. Letters indicate differences between groups. º p values were obtained from the Chi-squared test.

, no statistically significant differences were observed among the groups in terms of anthropometric, clinical, biochemical, or body composition variables. However, trends were noted in the group with high adherence to the EAT-Lancet diet, including lower weight, BMI, and muscle mass, alongside higher body fat levels.

Biochemical parameters showed a trend towards lower glucose, triglyceride, and total cholesterol levels in the high-adherence group. For glucose, this group presented levels of 88.79 ± 9.93 mg/dL, compared to 90.85 ± 7.71 mg/dL in the low-adherence group. Triglyceride levels were 82.71 ± 41.84 mg/dL in the high-adherence group and 98.85 ± 58.15 mg/dL in the low-adherence group. Total cholesterol levels were 153.39 ± 16.32 mg/dL in the high-adherence group, compared to 160.82 ± 36.68 mg/dL in the low-adherence group.

Blood pressure also tended to be lower in the high-adherence group. Conversely, this group exhibited lower physical activity levels. Despite these observed trends, no statistically significant differences were identified among the groups.

3.3. Nutrient Intake According to Adherence Level to the EAT-Lancet Diet

Table 4. Nutrient intake according to adherence level to the EAT-Lancet diet.

Variable	General Subsample		Adherence Level						p Value
			Low		Moderate		High
			n (%)		n (%)		n (%)
	n = 54		20 (37.00) ^		20 (37.00) ^		14 (26.00) ^
	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Energy (Kcal)	3296.02	1245.91	3494.36 a	1525.27	3201.09 a	1137.81	3148.29 a	968.67	0.8703
Fibre (g)	34.63	17.73	37.19 a	22.47	32.77 a	15.06	33.62 a	13.98	0.9597
Carbohydrates (g)	401.01	162.35	406.85 a	173.95	390.94 a	157.30	407.07 a	163.76	0.9597
Sugar (g)	152.76	102.42	147.07 a	66.60	127.60 a	75.43	196.81 a	157.64	0.3805
Protein (g)	133.37	59.85	153.74 a	75.25	131.74 a	51.86	106.62 a	31.36	0.0930
Lipids (g)	130.82	57.40	142.97 a	77.50	123.72 a	46.71	123.62 a	33.24	0.8561
Saturated fatty acids (g)	41.30	22.07	48.10 a	31.43	37.29 a	14.86	37.30 a	10.31	0.4009
Monounsaturated fatty acids (g)	37.57	18.38	40.95 a	25.04	35.48 a	13.71	35.73 a	12.60	0.9384
Polyunsaturated fatty acids (g)	25.00	16.47	24.48 a	17.66	21.81 a	12.71	30.31 a	19.18	0.3862
Cholesterol (mg)	574.20	333.76	712.18 a	409.69	528.62 ab	284.09	442.19 b	199.06	0.0158 *
Calcium (mg)	1559.99	670.65	1759.65 a	863.77	1516.73 a	508.87	1336.57 a	496.81	0.2664
Phosphorus (mg)	1868.01	844.85	2214.59 a	1098.95	1709.58 a	595.92	1599.23 a	569.82	0.1502
Iron (mg)	29.83	12.39	31.95 a	15.20	29.36 a	11.05	27.48 a	9.85	0.7200
Magnesium (mg)	524.15	241.56	570.37 a	293.15	501.28 a	205.19	490.78 a	213.25	0.7038
Sodium (mg)	4393.46	2739.18	4930.93 a	3968.11	3991.16 a	1881.36	4200.34 a	1292.13	0.6382
Potassium (mg)	4627.91	2192.74	5163.00 a	2751.99	4333.46 a	1859.35	4284.13 a	1667.12	0.6045
Zinc (mg)	16.76	6.85	18.99 a	8.71	15.80 a	5.36	14.93 a	5.08	0.3610
Selenium (mg)	55.99	26.36	61.67 a	27.49	57.99 a	29.67	45.01 a	16.02	0.2066
Vitamin A (µg RE)	1107.88	570.69	1223.46 a	571.23	1091.49 a	607.86	966.20 a	517.95	0.3260
Ascorbic acid (mg)	317.55	220.67	335.98 a	214.43	326.17 a	264.30	278.90 a	164.24	0.7860
Thiamine (mg)	2.77	1.47	2.89 a	1.36	2.89 a	1.82	2.41 a	1.04	0.4357
Riboflavin (mg)	3.35	1.82	3.84 a	1.95	3.36 a	1.95	2.66 a	1.21	0.0815
Niacin (mg)	26.36	11.85	30.05 a	13.37	26.03 a	11.91	21.58 a	7.66	0.1768
Pyridoxine (mg)	8.69	6.12	11.54 a	7.37	8.01 ab	4.79	5.61 b	4.03	0.0303 *
Folic acid (µg)	431.95	177.28	485.74 a	191.22	405.16 a	171.33	393.37 a	156.91	0.2617
Cobalamin (mg)	8.10	3.84	8.78 a	4.02	8.49 a	4.29	6.57 a	2.49	0.1998
Ethanol (g)	7.43	11.82	4.58 a	6.74	10.09 a	17.03	7.71 a	7.49	0.1914

Note: ^% from the total sample. SD = standard deviation. * p < 0.05 is considered statistically significant. p values were obtained from the Kruskal–Wallis test with Dunn’s post hoc analysis, as all data were non-normally distributed. Letters indicate differences between groups. Significant differences were highlighted in bold.

summarises the nutrient intake of the sample based on adherence level to the EAT-Lancet diet. A trend towards higher energy intake was observed in the low-adherence group (3494.36 ± 1525.27 Kcal/day) compared to the high-adherence group (3148.29 ± 968.67 Kcal/day). Although not statistically significant, lower protein intake was noted in the high-adherence group (106.62 ± 31.36 g/day) compared to the low-adherence (153.74 ± 75.25 g/day) and moderate-adherence groups (131.74 ± 51.86 g/day).

A statistically significant difference in cholesterol intake was observed across the adherence levels (p = 0.0158), with the low-adherence group consuming an average of 712.18 ± 409.69 mg/day, the moderate-adherence group consuming 528.62 ± 284.09 mg/day, and the high-adherence group consuming 442.19 ± 199.06 mg/day. Additionally, pyridoxine intake was significantly lower in the high-adherence group compared to the other groups (p = 0.0303).

3.4. Dietary Intake of the EAT-Lancet Diet Groups and Other Food Groups

Table 5. Intake of EAT-Lancet diet food groups according to adherence level to the EAT-Lancet diet.

Variable	EAT-Lancet Reference Intake			General Subsample		Adherence Level						p-Value
						Low		Moderate		High
						n (%)		n (%)		n (%)
	Suggested Intake	Possible Range		n = 54		20 (37.00) ^		20 (37.00) ^		14 (26.00) ^
	g/day	g/day		Mean	SD	Mean	SD	Mean	SD	Mean	SD
EAT-Lancet diet food groups 1				(g)	(g)	(g)	(g)	(g)	(g)	(g)	(g)
Whole grains	232.00	0.00%	60.00%	299.56	130.75	329.92 a	125.20	284.80 a	118.95	277.29 a	154.32	0.2239
Tubers and starchy vegetables	50.00	0.00	100.00	37.38	50.66	39.48 a	41.29	41.76 a	65.46	28.12 a	39.75	0.4139
Vegetables (all)	300.00	200.00	600.00	317.71	189.64	323.13 a	224.82	323.19 a	176.93	302.14 a	163.25	0.9115
Green vegetables	100.00	-	-	164.43	112.03	182.81 a	141.57	156.70 a	98.03	149.22 a	83.47	0.9545
Red and orange vegetables	100.00	-	-	91.15	70.93	91.41 a	69.71	96.95 a	82.39	82.49 a	57.94	0.9452
Other vegetables	100.00	-	-	62.13	49.84	48.91 a	37.89	69.54 a	54.11	70.43 a	57.65	0.4666
Fruits (all)	200.00	100.00	300.00	389.73	280.83	397.84 a	320.60	397.94 a	271.22	366.42 a	251.00	0.7535
Dairy foods	250.00	0.00	500.00	229.83	150.85	276.91 a	160.95	230.43 a	134.94	161.73 a	141.05	0.0674
Protein sources	84.00	0.00	211.00	229.95	139.65	292.43 a	166.65	224.89 ab	118.43	147.94 b	71.67	0.0012 *
Meats (protein sources without eggs)	71.00	0.00	186.00	165.52	99.85	209.29 a	122.38	166.22 ab	81.49	101.99 b	41.90	0.0016 *
Beef and lamb	7.00	0.00	14.00	44.54	28.27	49.44 a	29.49	45.74 a	28.45	35.84 a	26.15	0.3821
Pork	7.00	0.00	14.00	19.86	23.88	30.43 a	33.74	14.30 a	13.95	12.71 a	10.16	0.3951
Chicken and other poultry	29.00	0.00	58.00	58.48	53.96	84.05 a	65.65	48.62 a	46.40	36.01 a	26.94	0.0212 *
Eggs	13.00	0.00	25.00	64.44	55.27	83.14 a	58.87	58.67 a	53.12	45.94 a	48.04	0.0599
Fish	28.00	0.00	100.00	42.64	49.87	45.36 a	45.96	57.56 a	63.38	17.43 a	11.97	0.0497 *
Legumes	125.00	0.00	225.00	121.83	99.41	161.55 a	141.63	96.84 a	54.97	100.78 a	51.88	0.3986
Dry beans, lentils, peas, and chickpeas	50.00	0.00	100.00	98.22	95.49	127.31 a	140.28	73.41 a	47.87	92.10 a	52.68	0.5306
Soy foods	25.00	0.00	50.00	4.22	12.29	5.59 a	12.65	4.86 a	15.75	1.36 a	2.65	0.2276
Peanuts	25.00	0.00	75.00	6.00	15.69	9.13 a	24.75	5.07 a	7.07	2.86 a	2.80	0.9229
Tree nuts	25.00	-	-	13.39	34.07	19.52 a	44.22	13.50 a	34.23	4.47 a	4.32	0.1243
Added fats	51.80	20.00	91.80	78.23	90.43	83.23 a	82.09	83.64 a	124.30	63.38 a	24.30	0.6041
Saturated fats	6.80	0.00	6.80	41.30	22.07	48.10 a	31.43	37.29 a	14.86	37.30 a	10.31	0.4009
Unsaturated oils	40.00	20.00	80.00	16.90	18.25	12.99 a	11.15	17.85 a	23.86	21.11 a	17.30	0.3703
Dairy fats	0.00	0.00	0.00	16.02	64.11	13.79 a	35.19	26.90 a	99.89	3.64 a	3.13	0.6514
Lard or tallow	5.00	0.00	5.00	4.02	17.46	8.34 a	28.40	1.60 a	3.26	1.32 a	1.84	0.8615
Added sugars (all sweeteners)	31.00	0.00	31.00	152.76	102.42	147.07 a	66.60	127.60 a	75.43	196.81 a	157.64	0.3805

Note: ^% from the total sample. SD = standard deviation. p < 0.05 is considered statistically significant. p values were obtained from the Kruskal–Wallis test with Dunn’s post hoc analysis, as all data were non-normally distributed. Letters indicate differences between groups. ¹ Willett, W.; Rockström, J.; Loken, B.; Springmann, M.; Lang, T.; Vermeulen, S.; Garnett, T.; Tilman, D.; DeClerck, F.; Wood, A.; et al. Food in the Anthropocene: The EAT–Lancet Commission on Healthy Diets from Sustainable Food Systems. Lancet 2019, 393, 447–492. https://doi.org/10.1016/S0140-6736(18)31788-4 [10]. * Significant differences were highlighted in bold.

summarises the consumption of food groups according to adherence level to the EAT-Lancet diet. A statistically significant lower intake of chicken and other poultry was observed in the high-adherence group (36.01 ± 26.94 g/day) compared to the low- (84.05 ± 65.65 g/day) and moderate-adherence (48.62 ± 46.40 g/day) groups (p = 0.0212). Similarly, fish intake was lower in the high-adherence group (17.43 ± 11.97 g/day) compared to the low- (45.36 ± 45.96 g/day) and moderate-adherence (57.56 ± 63.38 g/day) groups (p = 0.0497).

Although not statistically significant, lower intakes of cereals, vegetables, and fruits were observed in the high-adherence group compared to the other groups. Similar trends were noted for foods of animal origin. Conversely, the high-adherence group showed higher sugar consumption and lower nut and peanut intake.

For non-EAT-Lancet food groups, the participants in the high-adherence group exhibited higher intakes of traditional Mexican foods, fatty cereals, and soft drinks. However, no statistically significant differences were identified among the groups for these items (

Table 6. Intake of non-EAT-Lancet diet food groups according to adherence level to the EAT-Lancet diet.

Variable	General Subsample		Adherence Level						p Value
			Low		Moderate		High
			n (%)		n (%)		n (%)
	n = 54		20 (37.00) ^		20 (37.00) ^		14 (26.00) ^
	Mean	SD	Mean	SD	Mean	SD	Mean	SD
Non-EAT-Lancet diet food groups
Fast food (g)	76.48	54.02	59.10 a	35.80	88.93 a	70.87	83.54 a	43.89	0.1968
Mexican food (g)	192.89	132.57	204.88 a	175.45	169.03 a	81.99	209.87 a	124.22	0.6338
Fatty cereals (g)	82.51	66.82	76.13 a	47.72	82.57 a	91.54	91.55 a	49.76	0.2573
Alcoholic beverages (g)	98.36	191.54	66.52 a	135.48	146.43 a	272.22	75.19 a	93.54	0.3333
Soft drinks (mL)	128.98	198.88	120.50 a	167.82	81.25 a	112.25	209.29 a	301.85	0.4066
Juices (mL)	34.37	39.23	37.06 a	43.98	29.63 a	37.16	37.28 a	37.06	0.7767
Coffee without milk (mL)	294.91	391.39	262.18 a	437.41	336.60 a	450.54	282.14 a	209.10	0.2155
Coffee without milk (mL)	24.06	48.26	30.98 a	62.99	23.99 a	46.53	14.29 a	18.87	0.5011
Fermented Mexican drinks (mL)	32.58	34.99	35.06 a	45.07	29.05 a	21.90	34.09 a	35.92	0.9644
Fresh fruit water (mL)	152.59	186.21	87.20 a	86.74	220.00 a	254.83	149.71 a	147.93	0.1233
Sport drinks (mL)	40.83	90.96	15.49 a	27.19	70.74 a	133.14	34.29 a	65.83	0.1324
Natural water (mL)	1413.77	916.07	1308.80 a	783.14	1514.78 a	1229.29	1419.43 a	532.60	0.8307
Artificial sweeteners (g)	0.45	1.30	0.58 a	1.72	0.46 a	1.22	0.26 a	0.58	0.9984

Note: ^% from the total sample. SD = standard deviation. p < 0.05 is considered statistically significant. p values were obtained from the Kruskal–Wallis test with Dunn’s post hoc analysis, as all data were non-normally distributed. Letters indicate differences between groups.

).

3.5. Gut Microbiota of the Total Subsample

The overall relative abundance of the selected gut bacteria in the total subsample is summarised in Figure 1. Dots represent the mean relative abundance for each bacterial group, and horizontal lines indicate the standard deviation (SD), reflecting the variability within the subsample. Firmicutes displayed the highest mean relative abundance, followed by Clostridium coccoides and Faecalibacterium prausnitzii, while Bilophila wadsworthia and Streptococcus thermophilus exhibited the lowest.

Statistical comparisons are represented by different letters next to each bacterial group. Groups sharing at least one letter are not significantly different from each other. For example, Firmicutes, Clostridium coccoides, and Faecalibacterium prausnitzii are statistically similar (group A). Bacteroidetes, Bifidobacterium, and Akkermansia muciniphila form a separate group (group B), also including Prevotella copri and Streptococcus thermophilus, which share overlap with other groups.

Lactobacillus is part of group C, differing from Firmicutes and Clostridium coccoides. Bilophila wadsworthia and Akkermansia muciniphila overlap in group D, indicating some similarity with bacteria from other groups like Prevotella and Faecalibacterium.

These distinctions suggest that while some bacteria exhibit clearly higher or lower abundances, others show intermediate levels with overlapping variability, as indicated by the shared letters in the statistical grouping.

3.6. Relative Abundance of Selected Gut Bacteria According to Adherence Level to the EAT-Lancet Diet

The relative abundance of the selected gut bacteria across the EAT-Lancet adherence groups is displayed in Figure 2. No statistically significant differences were observed among the groups for any of the bacteria analysed. The values were consistent and homogenous across all adherence levels, falling within a similar range.

Although no statistically significant differences were found between the groups for any of the bacteria, some trends were observed. For instance, in the high-adherence group, Firmicutes showed a tendency towards a higher relative abundance (0.182 ± 0.222) compared to the moderate- (0.143 ± 0.242) and low-adherence groups (0.148 ± 0.142). A similar trend was observed for Akkermansia muciniphila, with relative abundances of 0.013 ± 0.025, 0.011 ± 0.026, and 0.003 ± 0.008 in the high-, moderate-, and low-adherence groups, respectively. Likewise, Prevotella copri exhibited a similar pattern, with values of 0.065 ± 0.162, 0.063 ± 0.190, and 0.043 ± 0.128 across the high-, moderate-, and low-adherence groups, respectively. In contrast, Lactobacillus tended to be more abundant in the low- (0.0048 ± 0.0203) and moderate-adherence groups (0.000086 ± 0.00013) compared to the high-adherence group (0.000083 ± 0.00013).

3.7. Correlations Between Gut Microbiota, Adherence Level to the EAT-Lancet Diet, Metabolic, Anthropometric, Body Composition, Clinical Parameters, and Physical Activity

Correlations were performed between the relative abundance of the selected gut bacteria and adherence level to the EAT-Lancet diet, metabolic biomarkers, anthropometric, body composition and clinical parameters, and physical activity in the three adherence level groups and in the general subpopulation. Detailed correlation matrices are provided in Supplementary Materials S7–S10. No significant correlations were identified between the relative abundance of the selected gut bacteria and adherence level to the EAT-Lancet diet. However, the high-adherence group exhibited a statistically higher number of significant correlations with dietary intake, metabolic biomarkers, anthropometric measurements, body composition, clinical parameters, and physical activity variables compared to the other groups.

In the high-adherence group, specific correlations among bacterial species were observed. Positive correlations were identified between Lactobacillus and Bifidobacterium (rho = 0.5589; p = 0.0378), while negative correlations were found between Prevotella copri and Akkermansia muciniphila (rho = −0.7151; p = 0.0040) and Streptococcus thermophilus and Bifidobacterium (rho = −0.6821; p = 0.0072) (Table S8.1). Additional correlation matrices are detailed in Table S9.1 for the moderate-adherence group, Table S10.1 for the low-adherence group, and Table S11.1 for the general subpopulation.

Figure 3 highlights correlations in the high-adherence group involving adherence to the EAT-Lancet diet, metabolic markers, anthropometric and body composition data, clinical parameters, and physical activity. Among these, Prevotella copri showed negative correlations with body fat (rho = −0.578; p = 0.0304) and hip circumference (rho = −0.5396; p = 0.0464). Bacteroidetes negatively correlated with the waist-to-hip ratio (rho = −0.7143; p = 0.0041).

For metabolic biomarkers, triglycerides were positively associated with Lactobacillus (rho = 0.6492; p = 0.0120), while Bacteroidetes showed a negative correlation with total cholesterol (rho = −0.6088; p = 0.0209). Additionally, systolic blood pressure was negatively associated with Faecalibacterium prausnitzii (rho = −0.607; p = 0.0213), which also showed a positive correlation with the number of days per week of physical activity (rho = 0.5542; p = 0.0397).

In general, as adherence levels decreased, the number of correlations and their statistical significance also declined in the other groups. However, even in the low-adherence group, some correlations exceeded rho = 0.60. For instance, the waist-to-hip ratio and muscle mass positively correlated with Prevotella copri (rho = 0.4537; p < 0.05). Additionally, Akkermansia muciniphila showed significant negative correlations with BMI (rho = −0.4597; p < 0.05), metabolic age (rho = −0.5036; p < 0.05), and waist (rho = −0.4287; p < 0.05) and hip circumferences (rho = −0.5612; p < 0.05) (Table S10.2).

Correlations with metabolic biomarkers were also observed in the low- and moderate-adherence groups, as well as in the general subpopulation. Specifically, the low-adherence group displayed a positive correlation between triglycerides and Bilophila wadsworthia (rho = 0.4585; p < 0.05). In this group, a higher relative abundance of Prevotella copri was associated with higher systolic blood pressure (rho = 0.4918; p < 0.05) (Table S10.3). BMI negatively correlated with Akkermansia muciniphila (rho = −0.4597; p < 0.05), as did hip (rho = −0.5612; p < 0.05) and waist circumferences (rho = −0.4287; p < 0.05) (Table S10.2).

In the moderate-adherence group, significant correlations were identified between total and LDL cholesterol and Bifidobacterium (rho = −0.5008; rho = −0.4782; p < 0.05) as well as Streptococcus thermophilus (rho = −0.4962; rho = −0.4977; p < 0.05) (Table S9.3). BMI showed a negative correlation with Prevotella copri (rho = −0.4684; p < 0.05), while Bacteroidetes negatively correlated with hip circumference (rho = −0.4752; p < 0.05) (Table S9.2).

3.8. Correlations Between Gut Microbiota, Adherence Level to the EAT-Lancet Diet, Nutrients, and Food Group Intake

Correlations between the relative abundance of the selected gut bacteria and the intake of nutrients and food groups were analysed for the three adherence levels and the general subpopulation. Detailed correlation matrices are provided in Supplementary Materials S7–S10. Although no significant associations were found between the gut bacteria and adherence level to the EAT-Lancet diet, the high-adherence group exhibited statistically higher correlations than the other groups (Tables S8.1–S11.4).

Figure 4 illustrates the correlation matrix for the high-adherence group. In this group, beef and lamb intake negatively correlated with Bifidobacterium (rho = −0.5516; p = 0.0408) but positively correlated with Streptococcus thermophilus (rho = 0.5721; p = 0.0326). A negative correlation was observed between fish intake and Firmicutes (rho = −0.6249; p = 0.0169) as well as Akkermansia muciniphila (rho = −0.8458; p = 0.0001). Conversely, Prevotella copri positively correlated with fish intake (rho = 0.8119; p = 0.0004). Streptococcus thermophilus was positively associated with lard intake (rho = 0.5870; p = 0.0273), and coffee without added milk negatively correlated with Clostridium coccoides (rho = −0.5727; p = 0.0323).

As adherence levels decreased, the number and strength of correlations also diminished. In the moderate-adherence group, negative correlations were observed between Bifidobacterium and energy, carbohydrates, fibre, protein, and lipids, as well as with whole grains, total and green vegetables, dairy, fish, and fast food (rho > 0.4; p < 0.05). Detailed correlations for this group are listed in Table S9.4.

In the low-adherence group, Bacteroidetes negatively correlated with pork intake (rho = −0.4749; p = 0.0326), while Bifidobacterium positively correlated with poultry intake (rho = 0.4554; p < 0.05). Firmicutes showed a negative correlation with legume intake, particularly beans, lentils, peas, and chickpeas (rho = −0.5083; p < 0.05). All correlations for this group are detailed in Table S10.4.

4. Discussion

This study explored the association between the adherence level to the EAT-Lancet diet in a real-life habitual intake context and the relative abundance of selected gut bacteria in a Mexican subpopulation. The selected bacteria included Firmicutes, Bacteroidetes, Lactobacillus, Bifidobacterium, Akkermansia muciniphila, Faecalibacterium prausnitzii, Prevotella copri, Streptococcus thermophilus, Clostridium coccoides, and Bilophila wadsworthia. Our hypothesis was that higher adherence to the EAT-Lancet diet would be related to a higher relative abundance of metabolic-health-related bacteria, such as Lactobacillus, Bifidobacterium, Akkermansia muciniphila, and Faecalibacterium prausnitzii, and that there would be negative relationships with inflammation-associated bacteria, such as Bilophila wadsworthia.

Although no statistically significant differences in bacterial abundance were observed across adherence groups, interesting patterns emerged through the correlation analyses. However, it is essential to highlight that these findings represent associations rather than causal relationships; therefore, they should be interpreted as an initial effort to elucidate the potential interactions between the EAT-Lancet diet and the gut microbiota. Until controlled interventional studies are conducted, the direct effects of this dietary pattern on the gut microbiota cannot be firmly established.

However, specific food groups were interestingly correlated with bacteria, which have been previously associated with those intakes, especially in the group with high adherence to the EAT-Lancet diet. For example, beef and lamb intake negatively correlate with Bifidobacterium, a well-known genus for its probiotic effects [6]. Nevertheless, our findings identified a correlation between meat intake and Streptococcus thermophilus, a bacterium that has also been recognised for its probiotic effects related to dairy intake [46]. Nevertheless, in our study, dairy consumption was associated with Bifidobacterium only in the moderate-adherence group. This may be explained by the higher intake of probiotic-rich dairy products, such as yoghurt and cheese, in this group (230.43 ± 134.94 g/day) compared to the high-adherence group (161.73 ± 141.05 g/day) [46].

Similarly, intake of animal protein, such as fish, chicken, and beef, has been associated with abundance of bile-tolerant anaerobic bacteria, such as Bacteroides, Alistipes, and Bilophila [3]. However, our results indicate a negative correlation between meats and Bilophila wadsworthia. Nevertheless, the correlations were weak in this regard (rho < 0.33). In the specific case of fish, the high-adherence group showed a negative correlation between fish intake and Firmicutes and Akkermansia muciniphila. Conversely, Prevotella copri positively correlated with fish intake. A recent revision found that meat intake tends to decrease the Faecalibacterium genus [47]. However, in our study, no changes were identified in Faecalibacterium prausnitzii when comparing the adherence groups, even when meat intake was statistically lower in the high-adherence group. It is important to mention that the EAT-Lancet Commission recommends an average intake of meat of 71 g/day and a maximum consumption of 186 g/day [10]. Meanwhile, the high-adherence group consumed 101.99 ± 41.90 g/day, and the low-adherence group consumed 209.29 ± 122.38. However, those intakes did not reflect the changes in the gut microbiota.

Concerning vegetable protein foods, interesting correlations were found in soy consumption. In the low-adherence group, negative correlations were found with Prevotella copri. However, a positive correlation was found with Clostridium coccoides. Firmicutes negatively correlated with legume intake in the low-adherence group, particularly beans, lentils, peas, and chickpeas. Contrarily, the literature has associated vegetable protein intake with improved growth of the genera Bifidobacterium, Faecalibacterium, Clostridium, Eubacterium, and Roseburia, which are primary producers of butyrate and acetate [3]. Specifically, the intake of glycated pea proteins as plant-based proteins could increase the abundance of beneficial bacteria, such as Bifidobacterium and Lactobacillus, and decrease the abundance of Bacteroides fragilis and Clostridium perfringens [48].

Regarding fats, Streptococcus thermophilus was positively associated with lard intake in the high-adherence group. Studies have related saturated fat intake to bacteria such as Bilophila wadsworthia [6]. However, in our study, the bacterium associated with saturated fat-rich foods, such as lard, was Streptococcus thermophilus, specifically in the high-adherence group. In the case of monounsaturated fatty acids (palmitoleic, oleic, and eicosanoids), some studies associated their intake with Parabacteroides, Prevotella, Turicibacter genera, and the Enterobacteriaceae family [6]. However, in our study, monounsaturated fat-rich foods, such as tree nut intake, were only correlated with Bifidobacterium and Streptococcus thermophilus. Regarding polyunsaturated fat intake, their intake was only correlated with Faecalibacterium prausnitzii. Nevertheless, fish, which is high in this kind of fat, was correlated with Bifidobacterium.

The effects of coffee consumption on the gut microbiota have been a growing field of study. Some studies have associated its intake with Lawsonibacter saccharolytic, a butyrate-producing bacterium [4]. In our study, coffee without added milk negatively correlated with Clostridium coccoides. Coffee with milk was found to decrease Faecalibacterium prausnitzii. These findings oppose recent findings suggesting that moderate coffee intake (<4 cups per day) increases beneficial bacteria such as Bifidobacterium [49].

An interesting finding that opposes to most of the literature is related to the associations in the moderate-adherence group regarding some vegetable-origin foods. This group had negative correlations between Bifidobacterium and whole grains, total and green vegetables, and fibre intake. Nevertheless, energy, protein, dairy, fish, and fast food negatively correlated with these bacteria. The current literature generally relates fibre-rich foods to Bifidobacterium [3,6]. However, it is important to mention that these findings were obtained in the moderate-adherence group, for which intake of other food-groups, such as more meats and animal foods, could generate dysbiosis.

Although our findings provide interesting insights into this new field of study, it is important to consider that since our study was performed in a real-life habitual intake context, the specific effect of the EAT-Lancet diet cannot be established. In this regard, a recent study by Rehner [12] explored the effects of these dietary patterns under an intervention setting and compared their effects with other plant-based dietary patterns, such as vegetarians and vegans, for 2, 4, and 12 weeks. Although their study used massive sequencing for gut microbiota analysis, some of the bacteria identified coincide with the selected bacteria for this study.

Their finding indicates that the population adhering to the EAT-Lancet diet had constantly increased Bifidobacterium adolescentis. However, contrasting the planetary EAT-Lancet diet with vegans and vegetarians, no significant increases in specific probiotic bacteria, like Paraprevotella xylaniphila and Bacteroides clarus, were observed [12]. Among the bacteria analysed in our study, which coincide with the bacteria studied by Rehner [12], Prevotella copri stands out. This bacterium is capable of dietary fibre degradation, as it harbours vast genomic repertoires of carbohydrate-active enzymes [41]. Like Bifidobacterium adolescentis, switching to the EAT-Lancet diet might favour the growth of Prevotella copri. While SCFA-producing bacteria should benefit the host because of their anti-inflammatory and regulatory effects, Prevotella copri has also been correlated with the development of rheumatoid arthritis, although without conclusive evidence. An overgrowth of Prevotella copri also inhibits the growth of other beneficial microbiota [17].

Another study exploring the effects of this diet on the gut microbiome is the study by Deng et al. [13]. Their study was conducted in China in 2024 and evaluated the effects of the EAT-Lancet diet adherence level on the habitual intake of a prospective cohort (2008–2013). They identified Rothia mucilaginosa and Streptococcus sanguinis as being positively associated with the baseline EAT-Lancet score, and Bacteroides faecis, Ruminococcus torques, and Anaerostipes hadrus as being negatively associated with the baseline EAT-Lancet score [13]. Although the bacteria explored in that study do not agree with those studied in our work, some bacteria from similar genera, such as Clostridium, showed a trend towards high relative abundance in the high-adherence population. Similar trends were observed for health-related bacteria, such as Akkermansia muciniphila and Faecalibacterium prausnitzii.

Although the specific effects of the EAT-Lancet diet have been briefly explored, the number of studies examining the impact of plant-based diets on the gut microbiome is growing. Given the similar composition between these dietary patterns, comparable effects on gut microbial composition are expected. A noteworthy contribution in this field is the recent study by Fackelmann et al. [50], which reports clear distinctions between omnivorous and plant-based diets, highlighting associations between meat intake and bacteria such as Alistipes putredinis and Bilophila wadsworthia. In contrast, vegans showed greater abundance of butyrate-producing bacteria, including Lachnospiraceae, Butyricicoccus spp., and Roseburia hominis, while vegetarians were characterised by higher levels of Streptococcus thermophilus, likely due to dairy intake.

Although Bilophila wadsworthia did not differ significantly across the adherence groups in our study, other butyrate-associated bacteria, such as Akkermansia muciniphila, tended to be more abundant in the high-adherence group. Moreover, Prevotella copri, a bacterium strongly associated with plant-based diets, also exhibited a trend toward higher relative abundance in the high-adherence group compared to the others. An additional point of convergence with the findings of Fackelmann et al. [50] is the trend observed in Lactobacillus abundance: while their study reported higher levels in omnivores and vegetarians compared to vegans, our results similarly show greater relative abundance in the low- and moderate-adherence groups compared to the high-adherence group.

Although our paper’s main objective was to explore the association between the EAT-Lancet diet adherence level and the gut microbiota, interesting findings were obtained from the correlations between metabolic biomarkers, clinical indicators, and anthropometric and body composition variables. However, it is crucial to acknowledge that these correlations do not imply causality, and their biological significance remains unclear. The observed associations may be influenced by multiple confounding factors, including the gut environment, overall dietary fibre intake, host metabolism, and other unmeasured variables that could modulate microbial composition and metabolic responses [6]. Furthermore, while some bacterial taxa have been previously linked to metabolic health or inflammation, the extent to which these findings translate into meaningful physiological outcomes remains uncertain in the absence of direct mechanistic evidence [5].

Regarding metabolic biomarkers, triglycerides were positively associated with Lactobacillus, while Bacteroidetes showed a negative correlation with total cholesterol. Other studies have positively associated LDL cholesterol with Lachnospiraceae and Coriobacteriaceae and negatively with Bacteroidaceae and Bifidobacteriaceae [51]. Although the interaction between the gut microbiota and lipid profiles remains unclear, exploring the role of specific peptides and their implications in bile acid metabolism could be a key element for understanding these interactions [52].

Systolic blood pressure was negatively associated with Faecalibacterium prausnitzii, which also showed a positive correlation with the number of days per week of physical activity. In the low-adherence group, a higher relative abundance of Prevotella copri was associated with higher systolic blood pressure. These results agree with other research exploring the link between the gut microbiota and hypertension [53]. Further studies exploring the renin–angiotensin system could provide deeper insights since some bacteria associated with healthy or plant-based diets, such as Lactobacillus helveticus, produce antihypertensive peptides [54]. Additionally, other bacteria, such as Bifidobacterium longum 5022, have a maximum inhibitory potential of angiotensin-converting enzyme 1, thus modulating blood pressure [55]. However, this remains an area of study for future intervention research that evaluates the changes generated in the microbiota at the species or strain level, which will help confirm these effects.

Regarding body composition in the high-adherence group, Prevotella copri showed negative correlations with body fat and hip circumference. Bacteroidetes negatively correlated with the waist-to-hip ratio. Contrasting findings were recently reported [56], as fat mass and waist circumference correlate positively with Firmicutes and negatively with Bacteroidetes. In contrast, lean body and muscle mass positively correlate with Bacteroidetes.

In the low-adherence group, muscle mass was positively correlated with Prevotella copri. This finding aligns with a study on Polish athletes, which reported a higher abundance of Prevotella species in individuals engaged in regular physical activity. Given Prevotella copri’s role in carbohydrate fermentation and energy metabolism, its association with muscle mass warrants further investigation to determine whether it reflects dietary influences, gut microbiota adaptation to physical activity, or other metabolic interactions [57].

Regarding body weight, in our study, Akkermansia muciniphila showed significant negative correlations with BMI and waist and hip circumferences, contrasting with studies that have associated this bacterium with obesity [56,58]. These discrepancies may stem from differences in study populations, dietary patterns, or methodological approaches, highlighting the need for further research to clarify the role of Akkermansia muciniphila in body composition regulation.

Physical activity also showed interesting associations. It is worth mentioning that the strongest associations were also found in the high-adherence group. For instance, a positive correlation was found between Faecalibacterium prausnitzii and the number of days per week of physical activity. These results agree with a recent study exploring the effects of physical activity on the gut microbiota [59].

Even though the specific mechanisms have not been described yet, it has been proposed that producing some SCFAs, especially succinate (a substrate for intestinal gluconeogenesis), can improve glucose homeostasis and fatty acid oxidation, possibly through increased mitochondrial capacity [59,60,61].

These secondary findings offer valuable insights into the potential metabolic effects of the EAT-Lancet diet. However, as this study is cross-sectional and based on correlations, causal relationships cannot be established. Further interventional studies are needed to confirm whether adherence to this dietary pattern directly influences cardiometabolic biomarkers, body composition, and physical activity outcomes.

Despite the contribution of this work to the field, it is important to mention that this work presents limitations that must be considered to interpret our findings properly. The relatively small sample size (n = 54) may limit the generalizability of the findings, and the use of the Bonferroni correction, while minimising Type I errors, may have increased the risk of Type II errors, potentially overlooking relevant associations [45]. Future studies with larger sample sizes and alternative multiple comparison adjustments, such as the False Discovery Rate (FDR), are needed to validate these results.

Although qPCR is a widely used and cost-effective technique for bacterial quantification, metagenomic approaches, such as 16S rRNA sequencing or shotgun metagenomics, offer a broader perspective on gut microbiota composition and functionality. However, financial constraints prevented the implementation of metagenomic sequencing in this study. Despite this limitation, our qPCR-based approach enabled precise quantification of selected bacteria associated with dietary intake, providing valuable insights into the relationship between the EAT-Lancet diet and the gut microbiota.

Another point to consider is that this study focused primarily on species-level analysis but also included Firmicutes and Bacteroidota at the phylum level. This decision was based on the well-established predominance of these phyla in the gut microbiota, together accounting for approximately 90% of the total microbial community in healthy individuals, and their frequent use as indicators of broad compositional shifts. Moreover, this approach is consistent with previous studies employing targeted qPCR methodologies—such as Rodríguez-Lara et al. [18]—which have identified relevant associations between these phyla and dietary patterns in Mexican populations. Thus, while acknowledging the limited taxonomic resolution compared to sequencing, the inclusion of both phylum- and species-level taxa was guided by their functional relevance and grounded in the existing scientific literature.

Future research should incorporate advanced sequencing techniques to gain a more comprehensive understanding of this association in the Mexican population, along with α- and β-diversity analyses to assess microbiome composition and inter-individual variability. Additionally, metabolomic analyses could offer critical insights into microbial-derived metabolites and their interactions with human health outcomes, further elucidating the systemic effects of this dietary pattern [62,63].

5. Conclusions

The EAT-Lancet diet presents a promising dietary framework for achieving health and environmental sustainability goals. In this study, although no statistically significant differences in the relative abundance of selected gut bacteria were found across the adherence levels, specific and biologically plausible correlations emerged. Notably, higher adherence to the EAT-Lancet diet was associated with lower abundance of Bifidobacterium in relation to beef and lamb intake, and reduced levels of Akkermansia muciniphila were linked to fish consumption. Additionally, Prevotella copri and Faecalibacterium prausnitzii showed inverse correlations with body fat and blood pressure, respectively.

Furthermore, non-significant but biologically meaningful trends were observed in certain taxa. Akkermansia muciniphila and Prevotella copri tended to be more abundant in the high-adherence group, consistent with microbial profiles associated with plant-rich diets. In contrast, Lactobacillus showed higher relative abundance in the participants with low and moderate adherence, mirroring findings from previous studies comparing omnivorous and plant-based dietary patterns [51].

These findings suggest that even modest shifts toward a more sustainable dietary pattern may influence gut microbial composition in meaningful ways. However, the observational and targeted nature of the analysis—focusing on selected bacterial taxa—limits broader ecological interpretation and precludes causal inference. Future studies employing full microbiome sequencing and intervention designs are necessary to validate and expand upon these preliminary findings. Despite these limitations, our results contribute novel data from an underrepresented population and highlight the need for culturally sensitive strategies to promote sustainable diets with potential microbiome-related health benefits.