Abstract
Background
The EAT Lancet diet is increasingly recognized for its simultaneous benefits to human and planetary health. Its key components are known to exert anti-inflammatory and antiplatelet effects; however, the potential association between adherence to the EAT‑Lancet diet and incident venous thromboembolism (VTE), as well as the underlying biological mechanisms, remains unclear.
Methods
This prospective study involved 201,695 UK Biobank participants who were free of VTE at baseline, and integrated large-scale Olink plasma proteomics to identify diet-related molecular signatures. Adherence to the EAT-Lancet diet was quantified using two validated indices. Cox models were employed to evaluate associations between EAT-Lancet diet, plasma proteins, and incident VTE (including deep vein thrombosis [DVT] and pulmonary embolism [PE]). Mediation analyses quantified the role of plasma proteins.
Results
During a median follow-up of 13.77 years, participants in the highest adherence group of the Stubbendorff index showed lower risks of VTE (HR = 0.809, 95% CI: 0.750–0.874), DVT (HR = 0.856, 95% CI: 0.781–0.938), and PE (HR = 0.754, 95% CI: 0.678–0.838). Consistent protective associations were observed with the Knuppel index. Mediation analyses identified 94 shared plasma protein mediators, with mediation proportions ranging from − 4.77% to 8.45% of the association, among which FABP4, LEP, and ENPP6 were the top-ranked mediators. The overall proteomic signature mediated 59.19% (95% CI: 27.02%-92.15%) and 40.10% (95% CI: 14.53%-65.96%) of the associations between the Stubbendorff and Knuppel index and incident VTE, respectively.
Conclusions
Higher adherence to the EAT-Lancet diet is significantly associated with a reduced risk of incident VTE, DVT, and PE, with plasma proteins potentially mediating this effect.
Supplementary Information
The online version contains supplementary material available at https://doi.org/10.1186/s12916-026-04893-x.
Keywords: EAT-Lancet Diet, Venous Thromboembolism, Deep Vein Thrombosis, Pulmonary Embolism, Plasma Proteomics
Background
Venous thromboembolism (VTE), encompassing deep vein thrombosis (DVT) and pulmonary embolism (PE), represent the third most common cardiovascular disease globally, following myocardial infarction and stroke [1, 2]. Despite significant advancements in anticoagulant therapy, VTE remains associated with high recurrence rates, mortality, and enduring complications, notably post-thrombotic syndrome in survivors [3, 4]. While previous studies have well established that advanced age, surgery, immobilization, malignancy, and genetic susceptibility are major risk factors for VTE [5, 6], our understanding of modifiable lifestyle factors in primary VTE prevention, as well as the mechanisms linking environmental exposures to VTE pathogenesis, remains relatively limited.
Dietary modification is an easy yet high-impact intervention with broad health benefits. Prior research has demonstrated that the intake of ultra-processed foods is associated with an elevated risk of VTE [7], whereas adherence to a healthy plant-based diet [8], the Dietary Approaches to Stop Hypertension (DASH) diet [9] and anti-inflammatory diet index (AIDI) [10] is inversely associated with VTE risk. Unlike these dietary patterns, which prioritize human health alone, the EAT-Lancet Commission proposed the planetary health reference diet in 2019. This framework aims to simultaneously support human well-being and ecological sustainability, addressing the substantial environmental footprint associated with modern agricultural systems [11]. Characterized primarily by plant-based foods, this dietary plan emphasizes the consumption of whole grains, vegetables, fruits, nuts, and unsaturated fats, while substantially curtailing the consumption of added sugars, processed meats, and red meat [11]. Unlike rigid vegetarian regimens, the EAT-Lancet diet al.lows for modest amounts of animal-derived products and focuses on diverse plant consumption, ensuring it remains adaptable to various cultural contexts and local food supplies.
Prior epidemiological investigations have shown that following the EAT-Lancet guidelines correlates with a reduced risk of all-cause mortality [12], cardiovascular diseases [13], and type 2 diabetes [14, 15]. However, the potential link between EAT-Lancet diet adherence and the incidence of VTE and its subtypes remains largely unexplored. Given that VTE pathogenesis is rooted in oxidative stress and inflammation [16], and that key components of the EAT-Lancet diet exhibit anti-inflammatory and antiplatelet aggregation properties [17, 18], it is reasonable to hypothesize that high compliance with these dietary recommendations could lead to a decreased risk of incident VTE. Furthermore, the biological mechanisms underlying this potential association remain poorly understood. Plasma proteins are sensitive to environmental exposures and play pivotal roles in pathophysiological processes driving VTE [6, 19]. Investigating how EAT-Lancet diet adherence modulates plasma protein profiles may provide insights into potential molecular pathways linking this dietary pattern to VTE risk and may help identify candidate biomarkers for risk stratification and future investigation.
To address these knowledge gaps, this study aims to: [1] evaluate the longitudinal association between EAT-Lancet diet adherence and the risk of incident VTE and its subtypes; [2] utilize large-scale plasma proteomics data to systematically identify potential protein biomarkers related to EAT-Lancet diet and/or incident VTE; and [3] quantify mediating role of proteins in the diet-VTE association.
Methods
Study design and participants
Participants in this study were included from the UK Biobank, a prospective cohort study designed to investigate diverse health-related results. A detailed description of the study rationale and design is available in previous literature [20]. In summary, the baseline survey conducted between 2006 and 2010 involved over half a million UK residents aged 40 to 69.
In this study, we excluded participants without at least one valid 24-h dietary assessment (N = 291,197), those lost to follow-up (N = 568), those with implausible energy intake (≥ 18 MJ/day for women and ≥ 20 MJ/day for men; N = 3,159), and those with the history of VTE at baseline (N = 5,317). The final analytic sample for the primary analyses comprised 201,695 participants. For the mediation analysis evaluating the role of plasma proteins in the relationship between the EAT-Lancet diet and incident VTE, participants without available biomarker data were excluded (N = 183,703), resulting in a sample size of 17,992 (Additional file 1: Fig. S1). A schematic timeline of the study design is provided in Additional file 1: Fig. S2.
Assessment of EAT-Lancet diet index
To evaluate dietary consumption, we used the Oxford WebQ, an online, self-reported instrument designed to capture 24-hour dietary recall data [21]. Validation studies have previously confirmed the accuracy of this tool, showing that it yields nutrient and food group estimates comparable to those obtained via interviewer-led recalls [22]. Data collection occurred between 2009 and 2012, during which participants completed the questionnaire on up to five separate occasions, separated by intervals of 1 to 5 months. The distribution of participants contributing 1 to 5 dietary assessments is shown in Additional file 1: Table S1. To mitigate the impact of day-to-day fluctuations in eating habits, we computed the mean intake for individuals who provided data at multiple time points [23].
To quantify alignment with the EAT-Lancet Commission’s guidelines, we applied the scoring protocols from Knuppel et al. [24] and Stubbendorff et al. [25]. The Stubbendorff score (range: 0–42) assigns 0 to 3 points to 14 food items. Points are awarded for higher intake of seven healthy components (vegetables, fruits, whole grains, legumes, nuts, fish, and unsaturated oils) and lower consumption of seven restricted components (red/processed meats, poultry, eggs, dairy, potatoes, and sugar). The Knuppel score (range: 0–14) assesses these same categories using a binary scale, granting one point for each recommendation met. Full definitions are available in Additional file 1: Tables S2-S4.
Measurement of Olink proteomics
Proteomic data were derived from baseline samples (2006–2010). We employed the Olink Explore platform, which encompasses approximately 2,923 unique assays (UK Biobank category 1838) [26]. Following rigorous quality control and normalization procedures conducted by Olink facilities, data were provided as Normalized Protein Expression (NPX) values for every subject. The NPX unit serves as a log2-transformed metric for relative protein abundance. Detailed documentation regarding the QC workflow is available elsewhere [26].
After excluding participants who had more than 49.9% missing values and proteins with more than 10% missing values, retaining 45,175 out of 53,014 participants and 2,916 out of 2,923 proteins. The retained proteins included 731 from the cardiometabolic panel, 728 from the inflammation panel, 728 from the neurology panel, and 730 from the oncology panel. For proteins with less than 10% missing data, the k-nearest neighbors algorithm was employed to perform missing value imputation [27]. The specifics of the proteomic data could be seen in Additional file 1: Table S5.
Ascertainment of outcome
The primary study outcome was incident VTE, which includes both DVT and PE; these subtypes were also examined as secondary outcomes. Cases were identified from UK Biobank ‘First Occurrence’ fields, a resource that aggregates health information from multiple channels, including death registries, primary care logs, inpatient records, and self-reported data. VTE occurrences were defined according to the ICD (10th Revision), using codes I80-I82 (for DVT) and I26 (for PE) [28]. The observation period for each participant started on the date of the last diet questionnaire completion until the earliest of the following events: a VTE diagnosis, death, or the administrative censoring date of December 31, 2024. Additional file 1: Table S6 details the specific ICD codes used.
Assessment of covariates
The covariates for this study included sociodemographic factors (age, sex, ethnicity, education level, and Townsend Deprivation Index [TDI]) and lifestyle characteristics (body mass index [BMI], smoking status, alcohol consumption, physical activity, and total energy intake). Medical history covariates included diabetes, hypertension, cancer, hyperlipidemia, and aspirin use (see details in Additional file 2: Supplementary Methods [29–32]). To maximize data utilization, missing values were imputed with multiple imputations with chained equations (see details in Additional file 2: Supplementary Methods [33, 34]).
Statistical analysis
We investigated the relationship between EAT-Lancet diet compliance and VTE occurrence by employing the multivariable Cox proportional hazards models. We constructed three hierarchical models: Model 1 (unadjusted); Model 2 (adjusted for age, sex, ethnicity); and Model 3 (fully adjusted for education, TDI, BMI, smoking and alcohol habits, physical activity, aspirin use, and pre-existing conditions including diabetes, hypertension, hyperlipidemia, and cancer). Restricted cubic splines were applied to visualize non-linear trends [35].
Multivariable linear regression models (fully adjusted) were employed to evaluate the relationships between individual plasma proteins and the EAT-Lancet index. For the prospective analysis of plasma proteins and incident VTE risk, Cox models (fully adjusted) were utilized for each protein. Following this, Elastic Net regression analysis was implemented to select representative proteomic biomarkers from the pool of proteins significantly associated with EAT-Lancet diet. Elastic net models were fitted separately for the Stubbendorff index and the Knuppel index using the glmnet method implemented in the ‘caret’ package. This regularization technique, which combines Lasso and Ridge penalties, was chosen to minimize overfitting and handle collinearity [36]. We utilized 10-fold cross-validation to determine the optimal tuning parameters. Specifically, the optimal combination of α (predefined grid from 0.1 to 1.0 in increments of 0.1) and λ was selected according to the minimum cross-validated root mean squared error (RMSE). The final selected values were α = 0.1 and λ = 0.2287 for the Stubbendorff index model, and α = 0.1 and λ = 0.07383 for the Knuppel index model. The combined proteomic score was computed as the weighted aggregate of the proteins identified by Elastic Net. Finally, to explore whether the proteomic signature (or specific proteins) acted as a mediator in the association between the EAT-Lancet diet and incident VTE, we performed mediation analyses based on a hypothesized causal framework [37]. Specifically, individual proteins as well as proteomic signature were each evaluated separately as mediators in single-mediator models. The mediation proportion was determined by dividing the indirect effect by the total effect. The CIs were estimated via quasi Bayesian Monte Carlo method with 1000 simulations [38]. Furthermore, to explore the biological significance of the identified proteins, KEGG (Kyoto Encyclopedia of Genes and Genomes) functional enrichment analyses were performed.
To test result robustness, a set of sensitivity and subgroup analyses were performed, including excluding participants who completed only one occasion dietary recall, excluding incident VTE cases that occurred within the first two years, conducting a complete-case analysis by excluding participants with missing data, and assessing the associations between individual dietary components and VTE risk. Furthermore, we also examined effect modification by adding cross-product terms of EAT-Lancet diet indices with age, sex, socioeconomic status, and smoking status to the models, followed by corresponding stratified analyses. Given that several of the top mediating proteins are well-established adiposity-related proteins, we performed additional sensitivity analyses by further adjusting for waist circumference when evaluating the associations between the EAT-Lancet diet scores and VTE risk [39, 40]. We also repeated the mediation analyses for the top candidate proteins after further adjustment for waist circumference.
Statistical processing was performed in SAS 9.4 and R 4.5.1. For specific modeling tasks, we executed mediation analyses with the ‘mediation’ package and Elastic Net regression via the ‘glmnet’ and ‘caret’ package. A two-sided P-value of less than 0.05 indicated statistical significance for the relationships between the EAT-Lancet diet and VTE. To address multiple testing in proteomic studies, we considered a Benjamini-Hochberg FDR threshold of 0.05 as significant.
Results
Baseline characteristics
A total of 201,695 individuals were analyzed. Participants were classified into four groups using quartiles of the Stubbendorff EAT-Lancet index. The mean age of the overall population was 56.04 ± 7.95 years, 55.09% were females, and 95.73% were White. Table 1 summarizes that participants with elevated EAT-Lancet dietary scores exhibited a lower BMI and were less likely to be current smokers or daily alcohol consumers. They had higher levels of physical activity, lower total energy intake, and a higher rate of college education. Regarding health conditions, a lower prevalence of hypertension, hyperlipidemia, diabetes, and malignant tumors was noted in the high-adherence group compared to the lowest quartile. Consistency in these baseline characteristics was confirmed through stratification by the Knuppel EAT-Lancet index (Additional file 1: Table S7).
Table 1.
Baseline characteristics of participants according to Stubbendorff Eat-Lancet diet index categories
| Categories of the Stubbendorff EAT-Lancet diet score | ||||||
|---|---|---|---|---|---|---|
| Overall | Quartile 1 (< 22) | Quartile 2 (22–24) | Quartile 3 (24–27) | Quartile 4 (≥27) | ||
| (n = 201,695) | (n = 44,965) | (n = 56,064) | (n = 53,684) | (n = 46,982) | ||
| Age (mean [SD]) | 56.04 ± 7.95 | 55.60 ± 8.06 | 56.13 ± 7.96 | 56.23 ± 7.90 | 56.14 ± 7.85 | |
| Sex (n, %) | ||||||
| Female | 111,119 (55.09) | 20,924 (46.53) | 29,398 (52.44) | 30,843 (57.45) | 29,954 (63.76) | |
| Male | 90,576 (44.91) | 24,041 (53.47) | 26,666 (47.56) | 22,841 (42.55) | 17,028 (36.24) | |
| Ethnicity (n, %) | ||||||
| White | 193,083 (95.73) | 43,408 (96.54) | 53,971 (96.27) | 51,390 (95.73) | 44,314 (94.32) | |
| Others | 8,612 (4.27) | 1,557 (3.46) | 2,093 (3.73) | 2,294 (4.27) | 2,668 (5.68) | |
| Body mass index (kg/m2) (mean [SD]) | 26.91 ± 4.62 | 27.57 ± 4.71 | 27.22 ± 4.63 | 26.79 ± 4.57 | 26.05 ± 4.44 | |
| Physical activity (n, %) | ||||||
| Low | 36,610 (18.15) | 9,756 (21.70) | 10,649 (18.99) | 9,288 (17.30) | 6,917 (14.72) | |
| Moderate | 85,958 (42.62) | 18,941 (42.12) | 24,304 (43.35) | 22,914 (42.68) | 19,799 (42.14) | |
| High | 79,127 (39.23) | 16,268 (36.18) | 21,111 (37.66) | 21,482 (40.02) | 20,266 (43.14) | |
| Drinking (n, %) | ||||||
| Daily or almost daily | 46,149 (22.88) | 11,010 (24.49) | 12,888 (22.99) | 12,192 (22.71) | 10,059 (21.41) | |
| Three or four times a week | 50,584 (25.08) | 10,764 (23.94) | 14,117 (25.18) | 13,731 (25.58) | 11,972 (25.48) | |
| Once or twice a week | 50,380 (24.98) | 11,141 (24.78) | 14,133 (25.21) | 13,597 (25.33) | 11,509 (24.50) | |
| One to three times a month | 22,177 (11.00) | 5,052 (11.24) | 6,092 (10.87) | 5,825 (10.85) | 5,208 (11.09) | |
| Special occasions only | 19,876 (9.85) | 4,335 (9.64) | 5,504 (9.82) | 5,113 (9.52) | 4,924 (10.48) | |
| Never | 12,529 (6.21) | 2,663 (5.92) | 3,330 (5.94) | 3,226 (6.01) | 3,310 (7.05) | |
| Smoking (n, %) | ||||||
| Never | 114,393 (56.72) | 24,218 (53.86) | 31,601 (56.37) | 30,976 (57.70) | 27,598 (58.74) | |
| Previous | 71,522 (35.46) | 15,664 (34.84) | 19,963 (35.61) | 19,132 (35.64) | 16,763 (35.68) | |
| Current | 15,780 (7.82) | 5,083 (11.30) | 4,500 (8.03) | 3,576 (6.66) | 2,621 (5.58) | |
| Townsend deprivation index (mean [SD]) | -1.58 ± 2.87 | -1.53 ± 2.91 | -1.66 ± 2.83 | -1.64 ± 2.84 | -1.47 ± 2.91 | |
| Education (n, %) | ||||||
| ≥College | 86,329 (42.80) | 15,781 (35.10) | 22,577 (40.27) | 24,044 (44.79) | 23,927 (50.93) | |
| < College | 115,366 (57.20) | 29,184 (64.90) | 33,487 (59.73) | 29,640 (55.21) | 23,055 (49.07) | |
| Total energy intake (kcal/day) (mean [SD]) | 2,047.62 ± 568.69 | 2,067.61 ± 600.07 | 2,040.18 ± 575.08 | 2,034.70 ± 555.44 | 2,022.12 ± 544.03 | |
| History of aspirin use (n, %) | 24,133 (11.97) | 5,586 (12.42) | 6,939 (12.38) | 6,338 (11.81) | 5,270 (11.22) | |
| History of hypertension (n, %) | 48,756 (24.17) | 11,670 (25.95) | 14,022 (25.01) | 12,706 (23.67) | 10,358 (22.05) | |
| History of diabetes (n, %) | 8,343 (4.14) | 1,828 (4.07) | 2,323 (4.14) | 2,201 (4.10) | 1,894 (4.03) | |
| History of hyperlipidemia (n, %) | 22,218 (11.02) | 5,349 (11.90) | 6,304 (11.24) | 5,816 (10.83) | 4,749 (10.11) | |
| History of malignant tumor (n, %) | 14,217 (7.05) | 3,237 (7.20) | 3,980 (7.10) | 3,758 (7.00) | 3,242 (6.90) | |
Abbreviations: SD, standard deviation
Associations between EAT-Lancet diet index and risk of VTE
We observed linear associations of the Stubbendorff and Knuppel indices with incident VTE, DVT, and PE (Fig. 1) (P for nonlinearity > 0.05 for all). During a median (IQR) follow-up of 13.77 (13.36–14.58) years, we found that higher quartiles of the Stubbendorff EAT-Lancet index (Quartile 2-Quartile 4) correlated with lower risks of VTE, with respective HRs (95% CIs) of 0.868 (0.809–0.932), 0.863 (0.806–0.925) and 0.809 (0.750–0.874) compared with the lowest quartile in the model 3 (Table 2). The corresponding HRs (95% CIs) of higher adherence groups versus the lowest group of the Stubbendorff EAT-Lancet index for DVT were 0.891 (0.811–0.978), 0.873 (0.790–0.964) and 0.856 (0.781–0.938), and for PE were 0.855 (0.779–0.940), 0.824 (0.748–0.908) and 0.754 (0.678–0.838). Consistent findings were also obtained when analyzing the Knuppel EAT-Lancet diet index (Additional file 1: Table S8).
Fig. 1.
Dose-response relationships between EAT-Lancet diet indices and risk of venous thromboembolism and its subtypes. Restricted cubic spline (RCS) analyses illustrating the association between the Stubbendorff EAT-Lancet diet index and the Knuppel EAT-Lancet index with the hazard ratio (HR) for Venous Thromboembolism (VTE), Deep Vein Thrombosis (DVT), and Pulmonary Embolism (PE). The top row of plots displays the results for the Knuppel EAT-Lancet index, while the bottom row displays the results for the Stubbendorff EAT-Lancet diet index. The solid lines represent the estimated HR, and the shaded areas represent the 95% Confidence Interval (CI)
Table 2.
Associations between the Stubbendorff EAT-Lancet diet index and risks of VTE
| Cases/Total | HR (95% CI) | |||||
|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Model 3 | ||||
| Venous thromboembolism | ||||||
| Quartile 1 (< 22) | 1566/44,965 | Reference (1.000) | Reference (1.000) | Reference (1.000) | ||
| Quartile 2 (22–24) | 1665/56,064 | 0.847 (0.791, 0.908) | 0.838 (0.782, 0.898) | 0.868 (0.809, 0.932) | ||
| Quartile 3 (24–27) | 1554/53,684 | 0.823 (0.767, 0.883) | 0.820 (0.764, 0.880) | 0.863 (0.806, 0.925) | ||
| Quartile 4 (≥27) | 1199/46,982 | 0.724 (0.671, 0.780) | 0.739 (0.685, 0.797) | 0.809 (0.750, 0.874) | ||
| P-trend | < 0.001 | < 0.001 | < 0.001 | |||
| Continuous (per 1 score increase) | 5984/201,695 | 0.973 (0.966, 0.979) | 0.974 (0.968, 0.981) | 0.983 (0.976, 0.989) | ||
| Deep vein thrombosis | ||||||
| Quartile 1 (< 22) | 890/44,965 | Reference (1.000) | Reference (1.000) | Reference (1.000) | ||
| Quartile 2 (22–24) | 936/56,064 | 0.839 (0.766, 0.920) | 0.836 (0.763, 0.917) | 0.891 (0.811, 0.978) | ||
| Quartile 3 (24–27) | 903/53,684 | 0.844 (0.769, 0.925) | 0.851 (0.775, 0.934) | 0.873 (0.790, 0.964) | ||
| Quartile 4 (≥27) | 732/46,982 | 0.780 (0.707, 0.860) | 0.809 (0.733, 0.893) | 0.856 (0.781, 0.938) | ||
| P-trend | < 0.001 | 0.005 | ||||
| Continuous (per 1 score increase) | 3461/201,695 | 0.978 (0.970, 0.987) | 0.982 (0.973, 0.990) | 0.989 (0.980, 0.997) | ||
| Pulmonary embolism | ||||||
| Quartile 1 (< 22) | 856/44,965 | Reference (1.000) | Reference (1.000) | Reference (1.000) | ||
| Quartile 2 (22–24) | 899/56,064 | 0.838 (0.763, 0.920) | 0.823 (0.750, 0.904) | 0.855 (0.779, 0.940) | ||
| Quartile 3 (24–27) | 802/53,684 | 0.777 (0.706, 0.856) | 0.767 (0.696, 0.845) | 0.824 (0.748, 0.908) | ||
| Quartile 4 (≥27) | 604/46,982 | 0.667 (0.601, 0.740) | 0.673 (0.606, 0.748) | 0.754 (0.678, 0.838) | ||
| P-trend | < 0.001 | < 0.001 | 0.003 | |||
| Continuous (per 1 score increase) | 3161/201,695 | 0.966 (0.958, 0.975) | 0.966 (0.958, 0.975) | 0.977 (0.968, 0.986) | ||
HR, Hazard ratio; CI, Confidence interval
Model 1: unadjusted
Model 2: adjusted for age, sex, and ethnicity
Model 3: adjusted for Model 2 plus Townsend deprivation index, education level, BMI, smoking status, alcohol intake frequency, physical activity, total energy intake, history of aspirin use, diabetes, hypertension, hyperlipidemia and malignancy
Associations of the EAT-Lancet diet index with proteins
Following FDR correction, the proteomic analysis identified significant correlations between plasma proteins and the EAT-Lancet diet indices (Additional file 1: Tables S9-S10). Specifically, the Stubbendorff index was associated with 1584 proteins (236 positive and 1348 negative), while the Knuppel index showed associations with 847 proteins (206 positive and 641 negative), with 738 proteins (148 positive and 590 negative) shared between the two indices (Figs. 2 and 3A-B).Notable proteins displaying consistent positive associations with both dietary indices included key metabolic regulators such as adiponectin (ADIPOQ), lipoprotein lipase (LPL), and ghrelin (GHRL), as well as a distinct cluster of pancreatic digestive enzymes including pancreatic lipase (PNLIP), amylase (AMY1A/1B/1 C), and chymotrypsin C (CTRC). In contrast, proteins demonstrating consistent inverse associations were largely characterized by pro-inflammatory and thrombotic markers. These included key cytokines and adipokines such as interleukin-6 (IL6), tumor necrosis factor (TNF), and leptin (LEP), alongside hemostasis and matrix-remodeling proteins such as fibrinogen alpha chain (FGA), plasminogen (PLG), and matrix metalloproteinases (MMP9). Separately, Elastic Net regression selected 359 and 274 proteins to construct the Stubbendorff and Knuppel proteomic signature scores, respectively (Additional file 1: Tables S11-S12).
Fig. 2.
Venn diagrams illustrate the overlap of plasma proteins associated with the Stubbendorff and Knuppel indices. A all significant associations combined, (B) positive associations, and (C) negative associations
Fig. 3.
Volcano plots depicting the associations of plasma proteins with EAT-Lancet diet indices and incident VTE. Associations between plasma proteins and the Stubbendorff (A) or Knuppel (B) index. Hazard ratios calculated by Cox model for the risk of incident VTE associated with proteins identified in the Stubbendorff (C) and Knuppel (D) analyses. All model adjusted for age, sex, ethnicity, Townsend deprivation index, education level, BMI, smoking status, alcohol intake frequency, physical activity, total energy intake, history of aspirin use, diabetes, hypertension, hyperlipidemia and malignancy
Mediation analysis of plasma proteins
In the subset of participants with available Olink proteomic data, 553 incident VTE events were documented during a median follow-up of 13.76 years (IQR, 13.34–14.59). A total of 435 plasma proteins exhibited a significant correlation with incident VTE after FDR correction (Additional file 1: Table S13). Specifically, 253 and 168 proteins were simultaneously associated with the Stubbendorff and Knuppel indices, respectively, and incident VTE (Fig. 3C and D). Among these proteins, 190 and 116 of them exhibited significant mediation effects (Additional file 1: Tables S14-S15). There were 94 shared mediators for VTE between the two EAT-Lancet indices. The mediation proportions (95% CIs) of the shared mediators ranged from − 4.77% (-9.30%, -0.71%) to 8.45% (3.55%, 13.63%), among which fatty acid-binding protein 4 (FABP4), leptin (LEP), and ectonucleotide pyrophosphatase/phosphodiesterase 6 (ENPP6) were identified as the top-ranked mediators. Figure 4 shows the mediation proportions of mediators between the two indices and VTE. In addition, the overall proteomic signature scores mediated 59.19% (95% CI: 27.02% to 92.15%) and 40.10% (95% CI: 14.53% to 65.96%) of the associations between the Stubbendorff and Knuppel indices and incident VTE, respectively (Additional file 1: Table S16). Using KEGG methods, the significant mediators were mainly enriched into such pathways as Cytokine-cytokine receptor interaction, Viral protein interaction with cytokine and cytokine receptor, and Rheumatoid arthritis, which were consistently observed across the indices (Additional file 1: Tables S17-S18).
Fig. 4.
Mediation analysis of the top 20 significant proteins linking EAT-Lancet diet adherence to VTE risk. All model adjusted for age, sex, ethnicity, Townsend deprivation index, education level, BMI, smoking status, alcohol intake frequency, physical activity, total energy intake, history of aspirin use, diabetes, hypertension, hyperlipidemia and malignancy
Sensitivity and subgroup analyses
The inverse associations between the EAT-Lancet indices and the risk of incident VTE remained robust across a series of sensitivity analyses. These included excluding participants who completed the only one occasion dietary recall (Additional file 1: Table S19), excluding cases that occurred within the first 2 years of follow-up (Additional file 1: Table S20), conducting a complete-case analysis by excluding participants with missing data (Additional file 1: Table S21), and additionally adjusted for waist circumference (Additional file 1: Table S22). Analyses of individual dietary components generally supported the main findings, with specific components such as fruits and vegetables showing significant protective effects against VTE (Additional file 1: Tables S23-S24). Furthermore, we evaluated potential effect modifiers by stratifying participants according to age, sex, Townsend deprivation index, and smoking status. No significant interactions were observed, and the protective associations remained consistent across these subgroups (Additional file 1: Tables S25-S27). In additional sensitivity analyses adjusting for waist circumference, the mediating proportions of FABP4, LEP, and ENPP6 remained statistically significant for both the Stubbendorff and Knuppel EAT-Lancet diet scores, although some estimates were slightly attenuated (Additional file 1: Table S28).
Discussion
In this prospective cohort, we demonstrated that higher adherence to the EAT-Lancet diet as quantified by two distinct indices shows an inverse correlation with the risk of VTE, DVT, and PE. Additionally, we characterized a specific profile of plasma proteins associated with dietary adherence that is enriched in inflammatory, coagulation, and metabolic pathways. Notably, the overall proteomic signature mediated 59.19% and 40.10% of the relationships between the Stubbendorff and Knuppel EAT-Lancet indices and incident VTE, respectively.
The observed protective effect of the EAT-Lancet diet on VTE aligns with and extends prior research on dietary patterns and cardiovascular health. Previous epidemiological investigations have consistently shown that higher adherence to the EAT-Lancet diet is associated with reduced risks of cardiovascular diseases [41], and type 2 diabetes [14]. For instance, a longitudinal cohort investigation from ARIC study showed that individuals with the highest adherence to the EAT-Lancet reference diet had a 13% lower risk of total cardiovascular diseases (HR: 0.87; 95% CI: 0.79–0.96) [42]. These conditions share common pathophysiological pathways with VTE, including inflammation, oxidative stress, and metabolic dysfunction [43]. Our study is the first to demonstrate the protective role of the EAT-Lancet diet against VTE. Unlike exclusively health-focused patterns including the DASH and anti-inflammatory diets, the EAT-Lancet diet uniquely integrates human health with environmental sustainability [44]. Moreover, its allowance of moderate animal-source foods enhances cultural adaptability and long-term adherence compared to strict vegetarian diets [30], potentially making it a more practical dietary pattern for population-level VTE prevention.
The proteomic findings provide important insights into the biological mechanisms underlying this association. We found 253 proteins were concurrently associated with the Stubbendorff index and the risk of incident VTE, while 168 shared proteins were linked to the Knuppel index and VTE incidence. The larger number of shared proteins identified for the Stubbendorff index may largely reflect differences in score construction, particularly its wider multilevel scoring range compared with the binary Knuppel index. The overall pattern was highly consistent across both indices, suggesting that the association between the EAT-Lancet diet and VTE is linked to coordinated changes in inflammatory and immune-related proteomic pathways rather than to isolated biomarkers.
Pathway enrichment analyses further supported this interpretation. The most consistently enriched pathways across both dietary indices were cytokine-cytokine receptor interaction, viral protein interaction with cytokine and cytokine receptor, and rheumatoid arthritis. Rather than indicating disease-specific mechanisms per se, these pathway annotations collectively point to a broader network of immune and inflammatory signaling. In particular, the cytokine-cytokine receptor interaction pathway highlights the potential relevance of immunothrombotic processes, through which chronic low-grade inflammation may promote endothelial activation, platelet reactivity, coagulation factor expression, and impaired fibrinolytic balance, all of which are central to VTE development [45, 46].
The top mediating proteins also converged on this inflammatory-metabolic axis. FABP4 and LEP, which showed relatively strong mediation effects for both dietary indices, are closely linked to adiposity-related inflammation and vascular dysfunction. FABP4, an adipokine released from adipocytes and macrophages, has been implicated in endothelial dysfunction and vascular inflammatory responses [47, 48]. The EAT-Lancet diet’s emphasis on unsaturated plant oils and restriction of processed meats likely downregulates FABP4 expression, mitigating vascular injury. LEP, a marker of adiposity and direct pro-thrombotic factor, upregulates plasminogen activator inhibitor-1 (PAI-1) and enhances platelet aggregation [49]. Furthermore, LEP and FABP4 have previously been identified as BMI-related proteins in patients with established VTE, but the corresponding BMI-related proteomic signature did not materially mediate the obesity paradox for recurrent VTE or death [50]. Our findings supplement previous evidence by suggesting that these proteins may mediate the association between EAT-Lancet diet adherence and incident VTE, and that their role may extend beyond adiposity alone to broader metabolic and inflammatory pathways. ENPP6 may represent a more exploratory signal, potentially linking nutrient-related metabolic processes, including choline metabolism, to vascular or coagulation pathways [51]. We also identified overall proteomic signatures closely associated with the EAT-Lancet diet. This underscores that the EAT-Lancet diet acts through cumulative remodeling of the metabolic and inflammatory milieu, rather than a single pathway.
The findings have potential translational relevance for VTE prevention. Given the high recurrence rate and long-term complications of VTE, identifying modifiable lifestyle factors remains an important public health priority [52]. From a public health perspective, our observational findings suggest that greater adherence to the EAT-Lancet diet may be associated with a lower risk of VTE. Additionally, the identified proteins may offer insight into biological pathways underlying this association and may serve as candidates for future biomarker research. However, whether plasma proteomics can meaningfully predict future VTE risk remains an evolving and controversial question. Although previous studies suggest that circulating proteins may carry predictive information for incident VTE, the current evidence remains largely limited to biomarker discovery and association, and robust validation of predictive performance and clinical utility is still lacking [53–55]. Therefore, the predictive implications of these proteomic findings should be interpreted cautiously. In the present study, the proteomic analysis was designed primarily to explore potential biological mechanisms rather than to develop or validate a clinical prediction model. Independent external validation and future researches is needed before recommending this diet as a clinical intervention or considering these proteins for clinical risk stratification.
The primary strengths of this study include its prospective design, the large sample size from the UK Biobank, and the integration of high-throughput Olink proteomics, which allowed for a comprehensive exploration of molecular pathways. The use of two different EAT-Lancet diet scores further reinforces the robustness of our findings. However, some limitations must be acknowledged. First, as an observational study, residual confounding cannot be excluded. Although we adjusted for a wide range of covariates available in the UK Biobank, some established VTE risk factors were not fully captured. In addition, participants were not necessarily new adopters of the dietary pattern at baseline, individuals susceptible to VTE may already have experienced an event before dietary assessment, and dietary adherence may have changed during follow-up. Accordingly, the findings should be interpreted as associative rather than causal. Further independent external validation is needed before clinical or public health implementation. Second, dietary consumption and some covariates were evaluated using 24-hour recalls and self-reported questionnaires, which, despite being validated, may be subject to recall bias and measurement error; however, we mitigated this by averaging multiple assessments where available. Third, plasma proteins were measured only at baseline. Accordingly, the mediation findings should be interpreted as exploratory. Finally, participants in our study were predominantly white, healthier, and had a higher socioeconomic status than the general population in the UK [56]. Therefore, these findings may not be fully generalizable to the wider population.
Conclusions
This study suggests that higher adherence to the EAT-Lancet dietary pattern is associated with a lower risk of incident VTE, DVT, and PE. Plasma proteins may partly mediate these associations. These findings provide insight into potential biological mechanisms linking the EAT-Lancet dietary pattern to VTE risk.
Supplementary Information
Below is the link to the electronic supplementary material.
Supplementary Material 1: Additional file 1: Table S1–S29, Fig. S1–S2.
Supplementary Material 2: Additional file 2: Supplementary methods.
Acknowledgements
This research has been conducted using the UK Biobank Resource under Application Number 68136.
Abbreviations
- VTE
Venous thromboembolism
- DVT
Deep vein thrombosis
- PE
Pulmonary embolism
- ADIPOQ
Adiponectin
- AMY1A/1B/1C
Amylase
- CTRC
Chymotrypsin C
- ENPP6
Ectonucleotide pyrophosphatase/phosphodiesterase 6
- FABP4
Fatty acid-binding protein 4
- FGA
Fibrinogen alpha chain
- GHRL
Ghrelin
- IL6
Interleukin-6
- LEP
Leptin
- LPL
Lipoprotein lipase
- MMP9
Matrix metalloproteinases
- NPX
Normalized Protein Expression
- PAI-1
Plasminogen activator inhibitor-1
- PLG
Plasminogen
- PNLIP
Pancreatic lipase
- TNF
Tumor necrosis factor
- AIDI
Anti-inflammatory diet index
- BMI
Body mass index
- CI
Confidence Interval
- DASH
Dietary Approaches to Stop Hypertension
- FDR
False discovery rate
- HR
Hazard Ratio
- ICD
International Classification of Diseases
- KEGG
Kyoto Encyclopedia of Genes and Genomes
- TDI
Townsend Deprivation Index
Author contributions
QDH contributed to the conception and design of the study, drafted and revised the manuscript, and was responsible for the integrity and accuracy of the data analyses. MTS conceived and designed the study, contributed to result interpretation, and drafted the manuscript. YW conceived and designed the study and contributed to result interpretation. JZY contributed to result interpretation. YPS contributed to result interpretation and had full access to all the data used in this study. All authors read and approved the final manuscript.
Funding
No funding was received for this study.
Data availability
Access to the UK Biobank data used in this study is available through application to the UK Biobank in accordance with their established access procedure (https://www.ukbiobank.ac.uk/enable-your-research/apply-for-access).
Declarations
Ethics approval and consent to participate
The UK Biobank received ethics approval from the UK National Health Service’s National Research Ethics Service (ref. 11/NW/0382). Informed consent was obtained from all participants prior to their enrollment in the UK Biobank project.
Consent for publication
Not applicable.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Qida He and Mengtong Sun contributed equally to this work.
References
- 1.Kong Y, Tang W, Kang H, Guan Y, Li S, Cao X, et al. Linear and non-linear proteome-wide association studies provide novel insight into venous thromboembolism. Nat Commun. 2025;16(1):6517. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Lutsey PL, Zakai NA. Epidemiology and prevention of venous thromboembolism. Nat Rev Cardiol. 2023;20(4):248–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Kahn SR, Ginsberg JS. The post-thrombotic syndrome: current knowledge, controversies, and directions for future research. Blood Rev. 2002;16(3):155–65. [DOI] [PubMed] [Google Scholar]
- 4.Eichinger S, Lin M, Shi M, Grosso MA, Kyrle PA. Recurrent venous thromboembolism during anticoagulation with edoxaban or warfarin: A post hoc analysis of the Hokusai-VTE trial. Thromb Res. 2020;195:209–14. [DOI] [PubMed] [Google Scholar]
- 5.Khan F, Tritschler T, Kahn SR, Rodger MA. Venous thromboembolism. Lancet. 2021;398(10294):64–77. [DOI] [PubMed] [Google Scholar]
- 6.Pastori D, Cormaci VM, Marucci S, Franchino G, Del Sole F, Capozza A, et al. A Comprehensive Review of Risk Factors for Venous Thromboembolism: From Epidemiology to Pathophysiology. Int J Mol Sci. 2023;24(4):3169. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Yuan S, Chen J, Fu T, Li X, Bruzelius M, Akesson A, et al. Ultra-processed food intake and incident venous thromboembolism risk: Prospective cohort study. Clin Nutr. 2023;42(8):1268–75. [DOI] [PubMed] [Google Scholar]
- 8.Guo J, He YY, Suo XY, Liu Y, Xiao LL, Zhou YJ, et al. Healthy Plant-Based Diet, Genetic Predisposition, and the Risk of Incident Venous Thromboembolism. JACC Adv. 2024;3(12):101318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Li S, Jiang M, Guan Y, Cao X, Shao Z, Deng J, et al. Association between genetic risk and adherence to the Dietary Approaches to Stop Hypertension diet for developing venous thromboembolism. Res Pract Thromb Haemost. 2025;9(2):102731. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Yuan S, Bruzelius M, Damrauer SM, Hakansson N, Wolk A, Akesson A, et al. Anti-inflammatory diet and venous thromboembolism: Two prospective cohort studies. Nutr Metab Cardiovasc Dis. 2021;31(10):2831–8. [DOI] [PubMed] [Google Scholar]
- 11.Willett W, Rockstrom J, Loken B, Springmann M, Lang T, Vermeulen S, et al. Food in the Anthropocene: the EAT-Lancet Commission on healthy diets from sustainable food systems. Lancet. 2019;393(10170):447–92. [DOI] [PubMed] [Google Scholar]
- 12.Karavasiloglou N, Thompson AS, Pestoni G, Suter F, Papier K, Cassidy A, et al. Higher adherence to the EAT-Lancet reference diet is inversely associated with mortality in a UK population of cancer survivors. BMC Med. 2025;23(1):286. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Ye YX, Chen JX, Li Y, Lai YW, Lu Q, Xia PF, et al. Adherence to a planetary health diet, genetic susceptibility, and incident cardiovascular disease: a prospective cohort study from the UK Biobank. Am J Clin Nutr. 2024;120(3):648–55. [DOI] [PubMed] [Google Scholar]
- 14.Xu C, Cao Z, Yang H, Hou Y, Wang X, Wang Y. Association Between the EAT-Lancet Diet Pattern and Risk of Type 2 Diabetes: A Prospective Cohort Study. Front Nutr. 2021;8:784018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Zhang S, Zeng XF, Borne Y, Huo Z, Yan Y, Gu Y, et al. Adherence to the EAT-Lancet reference diet and risk of type 2 diabetes, metabolic dysfunction-associated steatotic liver disease, and their co-occurrence. Food Funct. 2025;16(17):6773–85. [DOI] [PubMed] [Google Scholar]
- 16.Rayes J, Brill A. Hot under the clot: venous thrombogenesis is an inflammatory process. Blood. 2024;144(5):477–89. [DOI] [PubMed] [Google Scholar]
- 17.Vita JA. Polyphenols and cardiovascular disease: effects on endothelial and platelet function. Am J Clin Nutr. 2005;81(1 Suppl):S292–7. [DOI] [PubMed] [Google Scholar]
- 18.Vazhappilly CG, Ansari SA, Al-Jaleeli R, Al-Azawi AM, Ramadan WS, Menon V, et al. Role of flavonoids in thrombotic, cardiovascular, and inflammatory diseases. Inflammopharmacology. 2019;27(5):863–9. [DOI] [PubMed] [Google Scholar]
- 19.Yuan S, Titova OE, Zhang K, Gou W, Schillemans T, Natarajan P, et al. Plasma protein and venous thromboembolism: prospective cohort and mendelian randomisation analyses. Br J Haematol. 2023;201(4):783–92. [DOI] [PubMed] [Google Scholar]
- 20.Allen NE, Lacey B, Lawlor DA, Pell JP, Gallacher J, Smeeth L, et al. Prospective study design and data analysis in UK Biobank. Sci Transl Med. 2024;16(729):eadf4428. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.He Q, Sun M, Zhao H, Sun N, Han Q, Feng Z, et al. Ultra-processed food consumption, mediating biomarkers, and risk of chronic obstructive pulmonary disease: a prospective cohort study in the UK Biobank. Food Funct. 2023;14(19):8785–96. [DOI] [PubMed] [Google Scholar]
- 22.Greenwood DC, Hardie LJ, Frost GS, Alwan NA, Bradbury KE, Carter M, et al. Validation of the Oxford WebQ Online 24-Hour Dietary Questionnaire Using Biomarkers. Am J Epidemiol. 2019;188(10):1858–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Cai YW, Gao JW, Wu MX, Xie YX, You S, Liao GH, et al. Adherence to EAT-Lancet diet, biological aging, and life expectancy in the UK Biobank: a cohort study. Am J Clin Nutr. 2025;122(1):29–38. [DOI] [PubMed] [Google Scholar]
- 24.Knuppel A, Papier K, Key TJ, Travis RC. EAT-Lancet score and major health outcomes: the EPIC-Oxford study. Lancet. 2019;394(10194):213–4. [DOI] [PubMed] [Google Scholar]
- 25.Stubbendorff A, Stern D, Ericson U, Sonestedt E, Hallstrom E, Borne Y, et al. A systematic evaluation of seven different scores representing the EAT-Lancet reference diet and mortality, stroke, and greenhouse gas emissions in three cohorts. Lancet Planet Health. 2024;8(6):e391–401. [DOI] [PubMed] [Google Scholar]
- 26.Eldjarn GH, Ferkingstad E, Lund SH, Helgason H, Magnusson OT, Gunnarsdottir K, et al. Large-scale plasma proteomics comparisons through genetics and disease associations. Nature. 2023;622(7982):348–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Goeminne LJE, Vladimirova A, Eames A, Tyshkovskiy A, Argentieri MA, Ying K, et al. Plasma protein-based organ-specific aging and mortality models unveil diseases as accelerated aging of organismal systems. Cell Metab. 2025;37(1):205–22. e6. [DOI] [PubMed] [Google Scholar]
- 28.Hu Z, Xu J, Shen R, Lin L, Su Y, Xie C, et al. Combination of Biological Aging and Genetic Susceptibility Helps Identifying At-Risk Population of Venous Thromboembolism: A Prospective Cohort Study of 394 041 Participants. Am J Hematol. 2025;100(4):575–83. [DOI] [PubMed] [Google Scholar]
- 29.Townsend P, Phillimore P, Beattie A. Health and deprivation. Inequality and the North. Revista cubana de higiene y epidemiología. 1997;35(1):48–50. [Google Scholar]
- 30.Feng J, Pan C, Gou Y, Yang X, Cheng S, Wei W, et al. EAT-Lancet Diet Modifies the Risk of Rheumatoid Arthritis Through Metabolomic Signature. Arthritis Rheumatol. 2026;78(1):68–79. [DOI] [PubMed] [Google Scholar]
- 31.Zhang Z, Huang Z, Xi Y, Han S, Ye X, Zhu H, et al. Association between adherence to the EAT-Lancet diet and risk of microvascular complications in type 2 diabetes: A cohort study. Diabetes Obes Metab. 2025;27(7):3858–68. [DOI] [PubMed] [Google Scholar]
- 32.Xu J, Ren Q, Su Y, Lin L, Shen R, Huang K, Frailty. Genetic Risk, and Long-Term Risk of Venous Thromboembolism: Insight From a UK Biobank Study. Am J Hematol. 2025;100(9):1523–32. [DOI] [PubMed] [Google Scholar]
- 33.Austin PC, White IR, Lee DS, van Buuren S. Missing Data in Clinical Research: A Tutorial on Multiple Imputation. Can J Cardiol. 2021;37(9):1322–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.White IR, Royston P, Wood AM. Multiple imputation using chained equations: Issues and guidance for practice. Stat Med. 2011;30(4):377–99. [DOI] [PubMed] [Google Scholar]
- 35.He Q, Sun M, Huang J, Wang Q, Adamkiewicz G, Shen Y, et al. Water hardness and digestive diseases: a large-scale population-based prospective cohort study. Int J Surg. 2026;112(1):922–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.Friedman J, Hastie T, Tibshirani R. Regularization Paths for Generalized Linear Models via Coordinate Descent. J Stat Softw. 2010;33(1):1–22. [PMC free article] [PubMed] [Google Scholar]
- 37.VanderWeele TJ. Mediation Analysis: A Practitioner’s Guide. Annu Rev Public Health. 2016;37:17–32. [DOI] [PubMed] [Google Scholar]
- 38.Tingley D, Yamamoto T, Hirose K, Keele L, Imai K. Mediation: R package for causal mediation analysis. J Stat Softw. 2014;59:1–38.26917999 [Google Scholar]
- 39.Ma H, Wang X, Heianza Y, Manson JE, Qi L. Proteomic Signature of BMI and Risk of Cardiovascular Disease. Clin Chem. 2024;70(12):1474–84. [DOI] [PubMed] [Google Scholar]
- 40.Zaghlool SB, Sharma S, Molnar M, Matias-Garcia PR, Elhadad MA, Waldenberger M, et al. Revealing the role of the human blood plasma proteome in obesity using genetic drivers. Nat Commun. 2021;12(1):1279. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Chen S, Ji C, Ji X, Wang C, Zhang R, Lu D, et al. Adherence to the EAT-Lancet Diet, Genetically Predicted Gut Butyrate and the Risk of Incident Cardiometabolic Multimorbidity: A Prospective Analysis. Mol Nutr Food Res. 2025;69(21):e70188. [DOI] [PubMed] [Google Scholar]
- 42.Yang J, Sullivan VK, Rebholz CM. Healthy Eating for the Planet, Cardiovascular Health, and Longevity: An Observational Study. J Am Heart Assoc. 2025;14(15):e040610. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Budnik I, Kumskova M, Chauhan AK. Metabolic pathways in deep vein thrombosis: a new frontier for therapeutic intervention. Blood. 2025;146(1):29–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Zhang S, Yan Y, Zeng XF, Gu Y, Wu H, Zhang Q, et al. Associations of the EAT-Lancet reference diet with metabolic dysfunction-associated steatotic liver disease and its severity: A multicohort study. Hepatology. 2025;81(5):1583–94. [DOI] [PubMed] [Google Scholar]
- 45.Koupenova M, Clancy L, Corkrey HA, Freedman JE. Circulating Platelets as Mediators of Immunity, Inflammation, and Thrombosis. Circ Res. 2018;122(2):337–51. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Stark K, Massberg S. Interplay between inflammation and thrombosis in cardiovascular pathology. Nat Rev Cardiol. 2021;18(9):666–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Couselo-Seijas M, Vazquez-Abuin X, Gomez-Lazaro M, Pereira L, Gomez AM, Caballero R, et al. FABP4 Enhances Lipidic and Fibrotic Cardiac Structural and Ca(2+) Dynamic Changes. Circ Arrhythm Electrophysiol. 2024;17(9):e012683. [DOI] [PubMed] [Google Scholar]
- 48.Furuhashi M. Fatty Acid-Binding Protein 4 in Cardiovascular and Metabolic Diseases. J Atheroscler Thromb. 2019;26(3):216–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 49.Singh P, Peterson TE, Barber KR, Kuniyoshi FS, Jensen A, Hoffmann M, et al. Leptin upregulates the expression of plasminogen activator inhibitor-1 in human vascular endothelial cells. Biochem Biophys Res Commun. 2010;392(1):47–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Ten Cate V, Koeck T, Prochaska J, Schulz A, Panova-Noeva M, Rapp S, et al. A targeted proteomics investigation of the obesity paradox in venous thromboembolism. Blood Adv. 2021;5(14):2909–18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Stewart AJ, Leong DTK, Farquharson C. PLA(2) and ENPP6 may act in concert to generate phosphocholine from the matrix vesicle membrane during skeletal mineralization. FASEB J. 2018;32(1):20–5. [DOI] [PubMed] [Google Scholar]
- 52.Stevens SM, Woller SC, Kreuziger LB, Bounameaux H, Doerschug K, Geersing GJ, et al. Antithrombotic Therapy for VTE Disease: Second Update of the CHEST Guideline and Expert Panel Report. Chest. 2021;160(6):e545–608. [DOI] [PubMed] [Google Scholar]
- 53.Edfors F, Iglesias MJ, Butler LM, Odeberg J. Proteomics in thrombosis research. Res Pract Thromb Haemost. 2022;6(3):e12706. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 54.Jensen SB, Hindberg K, Solomon T, Smith EN, Lapek JD Jr., Gonzalez DJ, et al. Discovery of novel plasma biomarkers for future incident venous thromboembolism by untargeted synchronous precursor selection mass spectrometry proteomics. J Thromb Haemost. 2018;16(9):1763–74. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Tang W, Li A, Austin TR, Braekkan SK, Nost TH, Li X, et al. Novel Plasma Proteomic Markers and Risk of Venous Thromboembolism. Circulation. 2026;153(11):810–25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56.Fry A, Littlejohns TJ, Sudlow C, Doherty N, Adamska L, Sprosen T, et al. Comparison of Sociodemographic and Health-Related Characteristics of UK Biobank Participants With Those of the General Population. Am J Epidemiol. 2017;186(9):1026–34. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Material 1: Additional file 1: Table S1–S29, Fig. S1–S2.
Supplementary Material 2: Additional file 2: Supplementary methods.
Data Availability Statement
Access to the UK Biobank data used in this study is available through application to the UK Biobank in accordance with their established access procedure (https://www.ukbiobank.ac.uk/enable-your-research/apply-for-access).