Sage Journals: Discover world-class research

Abstract

Background

Blood metabolites are crucial in various aspects of human health. However, current evidence regarding the role of circulating metabolites in various cardio-cerebrovascular diseases (CVDs) is limited. Mendelian randomisation (MR) can be used to provide information about these metabolites.

Objectives

Our primary aim was to investigate the causal relations between human blood metabolites and the risk of 10 CVDs.

Methods

We conducted a two-sample MR study. Data on 486 human blood metabolites were obtained from a genome-wide association study (GWAS) involving 7824 participants. Outcome data were sourced from the most recent large-scale GWAS meta-analysis by FinnGen (R9) on various CVDs. The inverse variance–weighted (IVW) model was used as the primary method for this analysis. Moreover, we performed sensitivity analyses, which included heterogeneity testing, horizontal pleiotropy testing and leave-one-out analysis, to evaluate the robustness and credibility of the findings.

Results

This study identified 36 known metabolites potentially associated with 10 CVDs, and sensitivity analyses showed no significant heterogeneity or pleiotropy.

Conclusions

Our study utilised systematic MR analyses and provides evidence for the causal relations between blood metabolites and CVDs. This sheds new light on the potential mechanisms underlying various CVDs and holds significant implications for screening, preventing and treating these conditions.

Keywords

blood metabolites cardiovascular diseases cerebrovascular diseases mendelian randomisation causality

Introduction

Cardio-cerebrovascular diseases (CVDs) remain the leading cause of death worldwide, accounting for 17.9 million deaths in 2019—32% of all global deaths—with 85% resulting from heart attacks and strokes, and with most occurring in low- and middle-income countries.¹ This mortality pattern has remained persistent.² The prevalence of CVDs increased markedly during the early twentieth century and plateaued in the 1980s in high-income regions.³ Globally, CVD-related deaths rose from 14.4 million to 17.5 million by 2005.⁴ The high incidence and mortality impose substantial disability and socioeconomic burden, underscoring the need for effective strategies in screening, prevention, and treatment. Established risk factors include elevated body mass index (BMI),^5,6 diabetes,^7–9 hyperlipidemia¹⁰ and smoking^11,12; however, research on metabolic alterations associated with CVDs remains limited.

Metabolomics is a rapidly evolving field that investigates the high-throughput characterisation of metabolites, encompassing metabolite composition across cell types, tissues, organs and organisms. Metabolomics consolidates endogenous small molecules to identify specific cellular biochemical fingerprints.¹³ This large-scale study provides a new method to explore the underlying mechanisms of diseases. Particularly, metabolic pathways and biomarkers associated with particular conditions are revealed by analysing and comparing changes in metabolites between healthy and diseased states. Human metabolomics investigations of CVDs have reinforced previous insights into the pathomechanisms of CVDs, implicating atherosclerosis, apoptosis, inflammation, oxidative stress and insulin resistance. Moreover, emerging research has highlighted that CVDs often appear as a downstream repercussion of metabolic dysregulation at the systemic or cellular levels and has revealed biomarkers and mechanisms of metabolic dysfunction in CVDs, potentially driving the dissection of disease mechanisms and the identification of novel therapeutic targets.¹⁴ Therefore, metabolomics provides a vantage point for understanding the mechanisms underlying CVDs and instigates innovative therapeutic strategies.

The potential relationships between circulating metabolic products and CVDs are significant and relevant. Although there are existing studies in this field,^15,16 many have been influenced by confounding factors, which weaken their conclusions and make further investigation both challenging and complex.

Mendelian randomisation (MR) is a widely utilised epidemiological investigative method that integrates summary data from genome-wide association studies (GWAS) by identifying single nucleotide polymorphisms (SNPs), thereby minimising the impact of confounding factors to infer causal effects between exposure factors and outcomes.

This study used a two-sample MR technique to assess the causal links between 486 human circulating metabolites and the risk of onset for 10 distinct CVDs to provide an understanding of the underpinnings of CVDs.

Materials and Methods

Study Design

Studies using MR should adhere to three fundamental postulates: 1) The genetic variants employed for measuring exposure must be linked with a genetic marker (blood metabolites). 2) No confounding factors should be present that would influence the relation between the genetic variants and the observed outcomes. 3) No direct connection should exist between the genetic variants and the outcomes that bypass the exposure route (Figure 1).

Figure 1.

The three basic assumptions of two-sample Mendelian MR analysis are illustrated by directed acyclic graphs.

GWAS Data for the Human Blood Metabolites

The GWAS data of the human blood metabolites originated from the most comprehensive blood metabolite genome-wide association estimate conducted so far by Shin et al¹⁷ The study involved 7824 adults from two European population studies and reported significant genome-wide associations for 145 metabolic loci and their biochemical links with 486 human blood metabolites. Specifically, 2.1 million SNPs were identified from 309 known and 177 unknown metabolites. According to the Kyoto Encyclopedia of Genes and Genomes (KEGG) database, the 309 known metabolites were classified into eight categories, such as cofactors and vitamins, energy, amino acids, carbohydrates, lipids, nucleotides, peptides and xenobiotics.

GWAS Data for the 10 Cardiovascular and CVDs

Outcome data (Table 1) on the CVDs, including atrial fibrillation and flutter (AF), atrioventricular block (AVB), paroxysmal tachycardia (PT), abdominal aortic aneurysm (AAA), myocardial infarction (MI), pulmonary embolism (PE), transient ischaemic attack (TIA), angina pectoris (AP), intracerebral haemmorrhage (ICH) and death due to cardiac causes (DCC)^18–20 were derived from the FinnGen Data Freeze 9 (R9) summary-level findings, publicly disclosed on 11 May 2023.²⁰ Launched in autumn 2017 in Finland, the FinnGen study is a pioneering research project that merges genomic information with electronic health records. This unique international collaborative effort is the most extensive in its domain and aims to deepen our understanding of various health-related concerns. The latest data collection encompassed 377 277 individuals (210 870 women and 166 407 men), encompassing 20 175 454 genetic variants and 2272 distinct health outcomes.

Table 1.

Characteristics of the Summary Datasets for the Outcomes.

Phenotypes	Data Source	Sample Size (Case \| Control)	Ancestry	Covariates	ID	Link
AF	FinnGen	45 766 \| 191 924	European	sex, age, 10 PCs, genotyping batch	Finngen_R9_I9_AF	finngen_results
AVB		5536 \| 286 109			finngen_R9_I9_AVBLOCK
PT		9604 \| 191 924			finngen_R9_I9_PAROXTAC
AAA		3548 \| 349 539			finngen_R9_I9_ABAORTANEUR
MI		24 185 \| 313 400			finngen_R9_I9_MI_STRIC
PE		9243 \| 367 108			finngen_R9_I9_PULMEMB
TIA		18 398 \| 342 294			finngen_R9_I9_TIA
AP		34 456 \| 313 400			finngen_R9_I9_ANGINA
ICH		3749 \| 339 914			finngen_R9_I9_ICH
DCC		19 295 \| 357 982			finngen_R9_I9_K_CARDIAC

Abbreviations: PCs, principal components.

Selection of Instrumental Variables

A significance threshold of p < 1 × 10⁻⁵ was adopted to derive instrumental variables (IVs) from a pool of 486 metabolites due to the limited sample sizes of the metabolites. We utilised the 1000 Genomes Project as the reference cohort, setting a linkage disequilibrium threshold of R² < 0.1 with a physical proximity constraint of 500 kilobases.²¹ Additionally, the F-statistic was computed for each metabolite to evaluate the robustness of the IVs. The specific calculations were as follows²²: $F = \frac{(N - 1 - k) \times R^{2}}{k \times (1 - R^{2})}$ $R^{2} = \frac{2 \times E A F \times β^{2} \times (1 - E A F)}{2 \times E A F \times β^{2} \times (1 - E A F) + 2 \times E A F \times S E {(β)}^{2} \times (1 - E A F) \times N}$

F, F-statistics; N, the sample size of the GWAS for the association between SNPs and blood metabolites; k, the numbers of IV; R², the proportion of variance in blood metabolites accounted for by each IV; EAF, the effect allele frequency; β, the effect size of the SNP on the blood metabolites; SE(β)² the standard error of the genetic effect.

MR Analysis

To investigate the potential causal relations between human blood metabolites and the risk of CVDs, we employed advanced statistical methods such as inverse-variance weighting (IVW), MR-Egger regression, weighted median (WM) and a simple mode approach. The primary method utilised was IVW, which integrates Wald ratios (or ratio estimates) in an inverse variance weighted meta-analysis. This approach accounted for the heterogeneity among the studies included in the analysis and comprehensively evaluated the causal relations between blood metabolites and CVD risk.²³ Additionally, we used MR-Egger regression, which combines the Wald ratio (or ratio estimates) into a meta-regression (with an intercept and slope) to estimate the causal effect adjusted for any directional pleiotropy.^24,25 Subsequently, we utilised the WM method to calculate the 50^th percentile of an empirical density function based on either unweighted or inverse variance–weighted estimates of the Wald ratio. This method provides an assessment of the causal effect in the presence of potential outliers or heterogeneity, thereby enhancing the validity of our findings.²⁶ Moreover, a simple mode method was incorporated. Although not as statistically robust as the IVW method, the simple mode method adds robustness in cases of pleiotropy, which refers to a single genetic variant influencing multiple traits or outcomes.²⁷ The simple mode method helps to address potential bias arising from pleiotropy, thereby enhancing the reliability of the results. Thus, we aimed to provide a comprehensive assessment of the causal relations between blood metabolites and the risk of CVDs, thereby enhancing our understanding of the potential influence of blood metabolites on the development of CVDs.

Sensitivity Analysis

We conducted extensive sensitivity analyses to ensure the reliability and robustness of our findings. These investigations examined the potential heterogeneity and identified invalid IVs in the IVW model using various methods, including the Cochran's Q test.²⁸ We used the MR-Egger regression test²⁹ to assess horizontal pleiotropy^30,31 and the leave-one-out test to evaluate the effect of outlier SNPs on causal estimates. Moreover, considering the potential invalidity of the SNPs due to pleiotropy, particularly when the assessment method indicated balanced pleiotropy, we implemented the Mendelian randomisation pleiotropy residual sum and outlier (MR-PRESSO) strategy. This approach autonomously identifies outliers, thereby ensuring the accuracy of the analysis.³² To better understand and recognise the validity of these analyses, we visually presented the result indicators, such as MR scatter plots showing the estimated impact of exposure on outcomes, funnel plots illustrating the heterogeneity of estimates and forest plots displaying the variation in the relation between the exposure and outcome of each SNP. These sensitivity analyses reinforced the reliability of our findings and strengthened our understanding of the association between human blood metabolites and CVD risks.

Pathway and Enrichment Analysis

We analysed the pathways and enrichment of metabolites associated with the outcomes using MetaboAnalyst 5.0. Initially, we identified these metabolites using the Human Metabolome Database (HMDB). Subsequently, we employed enrichment analysis and pathway analysis features in the Annotated Features section of this tool. We compiled various sets of metabolites and their related pathways associated with CVDs by referencing the Small Molecular Pathways Database (SMPDB)³³ and Kyoto Encyclopedia of Genes and Genomes (KEGG)³⁴ database. Furthermore, multiple tests utilised the false discovery rate (FDR) correction to mitigate false positives. A result was considered significant if the FDR for the metabolic pathway and enrichment analysis was < 0.05.

Statistical Analysis

All statistical analyses were performed using the TwoSampleMR (version 0.5.6) and MR-PRESSO (version 1.0) packages in R software (version 4.3.2). We applied the Bonferroni correction to account for multiple tests across 10 different outcomes, resulting in a significance threshold of 0.005. A two-sided P-value < .05 was considered significant for all other analyses.

Results

IV Information

A total of 10 538 SNPs were significantly associated with 486 metabolites, with the number of chosen IVs for these metabolites varying between 2 and 493. Notably, the minimum F-statistic for the validity tests was >10 (the lowest was 17.64), indicating that all SNPs had sufficient validity. These details are presented in Supplementary Table S1.

MR Analysis Results

We excluded 177 unknown metabolites to better elucidate the metabolic changes. Most of the remaining 309 known blood metabolites were not associated with CVDs. The causal connections between these metabolites and 10 distinct outcomes are demonstrated in Figures 2A–J and 3. Supplementary Table S2 provides a comprehensive analysis of the entire MR study.

Figure 2.

The MR analysis results show the relationship between blood metabolites and various CVDs: (A) AF, (B) AVB, (C) PT, (D) AAA, (E) MI, (F) PE, (G) TIA, (H) AP, (I) ICH, (J) DCC.

Figure 3.

Volcano plot showing the causal estimates of 486 metabolites on CVDs in the primary analyses with inverse-variance weighted method. IVW. (A) AF, (B) AVB, (C) PT, (D) AAA, (E) MI, (F) PE, (G) TIA, (H) AP, (I) ICH, (J) DCC.

Notably, we identified 53 positive causal relations, of which 36 were with known metabolites based on different outcomes. Among these, 14 were from the lipidmetabolic pathway, eight from the amino acid metabolic pathway, four from the carbohydrate metabolic pathway, three from the cofactors and vitamins metabolic pathway, three from the energy metabolic pathway, two from the peptide metabolic pathway, one from the nucleotide metabolic pathway and one from the xenobiotic metabolic pathway (Figures 4 and 5). They were as follows: First for AF, 2-aminooctanoic acid [odds ratio (OR) = 0.70, 95% confidence intervals (CI) 0.58–0.82], isovalerylcarnitine (OR = 1.50, 95% CI = 1.18-1.92), N-acetylornithine (OR = 0.86, 95% CI = 0.78-0.93), 3 dehydrocarnitine (OR = 1.67, 95% CI = 1.18-2.36), butyrylcarnitine (OR = 1.14, 95% CI = 1.08-1.24), dodecanedioate (OR = 1.55, 95% CI = 1.18-2.04), tetradecanedioate (OR = 0.74, 95% CI = 0.68-0.84), inosine (OR = 1.11, 95% CI = 1.08-1.18); for AVB, 4-androsten-3beta,17beta-diol disulfate 1 (OR = 0.76, 95% CI = 0.68-0.89); for PT, kynurenine (OR = 1.92, 95% CI = 1.28-2.95), dodecanedioate (OR = 1.96, 95% CI = 1.28-3.07); for AAA, 5alpha-androstan-3beta,17beta-diol disulfate (OR = 1.36, 95% CI = 1.18-1.68), epiandrosterone sulfate (OR = 1.35, 95% CI = 1.18-1.66); then for MI, creatine (OR = 0.65, 95% CI = 0.58-0.86), glycine (OR = 0.70, 95% CI = 0.58-0.87), oleoylcarnitine (OR = 1.74, 95% CI = 1.182,57),palmitoleate (16:1n7) (OR = 0.49, 95% CI = 0.38-0.80), tetradecanedioate (OR = 0.79, 95% CI = 0.68-0.92), ADpSGEGDFXAEGGGVR (OR = 1.53, 95% CI = 1.18-2.05); for PE, 1,5-anhydroglucitol (1,5-AG) (OR = 1.96, 95% CI = 1.39-2.77), mannitol (OR = 0.78, 95% CI = 0.67-0.92), pyridoxate (OR = 0.60, 95% CI = 0.44-0.82), 1-arachidonoylglycerophosphocholine (OR = 1.75, 95% CI = 1.25-2.44), 1-linoleoylglycerophosphoethanolamine (OR = 0.50, 95% CI = 0.33-0.74), arachidonate (20:4n6) (OR = 2.32, 95% CI = 1.62-3.34); for TIA, 2-aminooctanoic acid (OR = 0.80, 95% CI = 0.68-0.93), cysteine-glutathione disulfide (OR = 0.73, 95% CI = 0.59-0.91), citrate (OR = 0.54, 95% CI = 0.37-0.77), succinylcarnitine (OR = 0.64, 95% CI = 0.48-0.86), ibuprofen (OR = 1.05, 95% CI = 1.02-1.08); for AP, threonate (OR = 0.80, 95% CI = 0.69-0.93), 1-palmitoleoylglycerophosphocholine (OR = 0.61, 95% CI = 0.44-0.85), adrenate (22:4n6) (OR = 0.61, 95% CI = 0.47-0.80), oleoylcarnitine (OR = 1.66, 95% CI = 1.18-2.33), palmitoleate (16:1n7) (OR = 0.58, 95% CI = 0.43-0.78), tetradecanedioate (OR = 0.82, 95% CI = 0.72-0.93); for ICH, mannose (OR = 0.39, 95% CI = 0.22-0.69), pyroglutamylglycine (OR = 2.13, 95% CI = 1.28-3.53); finally for DCC, phenol sulfate (OR = 1.61, 95% CI = 1.26-2.05), threitol (OR = 0.61, 95% CI = 0.46-0.82), o-methylascorbate (OR = 0.71, 95% CI = 0.56-0.90), phosphate (OR = 0.15, 95% CI = 0.06-0.38). The scatter plot of MR analysis shows each causal relationship in more detail (Supplementary Figure S1). We observed that different CVD phenotypes shared common causal metabolites. For example, dodecanedioate was causally linked with AF and PT, and tetradecanedioate was associated with AF, MI and AP, among others. Interestingly, these identical metabolites demonstrated consistent causal directions across different outcomes,albeit with varying degrees of effect.

Figure 4.

Mr associations of known metabolites on the risk of the 10 phenotypes of CVDs.

Figure 5.

The positive causal effects of blood metabolites on a variety of CVDs.

Evaluation of the Reliability and Stability of the Results

We performed sensitivity analyses, including heterogeneity and horizontal pleiotropy tests, to assess the reliability and stability of the results. First, Cochran's Q and MR-Egger regression tests were conducted to examine the heterogeneity and horizontal pleiotropy, respectively, of each previously mentioned positive finding (Table 2). No indication of horizontal pleiotropy was observed in the results. However, a few of these MR analyses were heterogenous. The primary analytical method used was IVW, which effectively avoids the impact of heterogeneity on causal effect estimate. The statistics of all IVs we selected were > 10, so the adverse effects of heterogeneity were ignored. Furthermore, MR-PRESSO tests identified some potential outliers; however, after adjusting for these outliers, the effect on the results was minimal, as the corrected P-value remained consistent with the raw P-value < .05 (Supplementary Table S3). Additionally, the funnel plot exhibited global symmetry, further reinforcing the validity of the results. The leave-one-out analysis and forest plots bolstered the reliability of the initial analytical conclusions (Supplementary Figures S2–S4).

Table 2.

Heterogeneity and Horizontal Pleiotropy Testing of Positive MR Results.

ID	nSNP	Heterogeneity Test			Pleiotropy Test			ID	nSNP	Heterogeneity Test			Pleiotropy Test
		IVW			MR-Egger					IVW			MR-Egger
		Cochran's Q	Q df	P Value	Intercept	SE	P value			Cochran's Q	Q df	P Value	Intercept	SE	P value
AF								PE
M01123	13	10.82	12	5.44E-01	7.45E-04	5.58E-03	8.96E-01	M01110	30	24.04	29	7.27E-01	−6.76E-04	6.40E-03	9.17E-01
M15630	25	18.72	24	7.67E-01	4.76E-03	3.74E-03	2.17E-01	M15335	13	8.46	12	7.48E-01	5.38E-03	1.68E-02	7.54E-01
M19364	20	21.39	19	3.16E-01	2.03E-03	7.10E-03	7.78E-01	M20675	37	46.21	36	1.19E-01	1.08E-02	7.93E-03	1.83E-01
M32388	7	3.29	6	7.72E-01	−2.51E-02	1.70E-02	2.01E-01	M31555	12	4.03	11	9.69E-01	−1.59E-03	9.99E-03	8.77E-01
M32412	51	87.00	50	9.23E-04	1.91E-03	4.03E-03	6.38E-01	M32635	14	10.80	13	6.27E-01	2.02E-02	1.34E-02	1.56E-01
M32654	29	43.21	28	3.32E-02	−1.02E-02	8.84E-03	2.58E-01	M33137	15	51.15	14	3.91E-06	−1.17E-02	2.88E-02	6.92E-01
M32808	16	40.18	15	4.27E-04	−1.12E-02	1.01E-02	2.85E-01	M33228	24	17.17	23	8.01E-01	−1.25E-02	6.90E-03	8.31E-02
M32855	23	27.41	22	1.96E-01	6.41E-03	6.05E-03	3.01E-01	TIA
M33751	28	30.32	27	3.00E-01	5.57E-03	4.19E-03	1.95E-01	M01564	49	51.82	48	3.27E-01	8.53E-03	5.45E-03	1.24E-01
M34040	20	25.48	19	1.45E-01	1.29E-03	5.67E-03	8.22E-01	M17799	127	134.40	126	2.88E-01	−3.54E-03	4.49E-03	4.33E-01
M34407	25	45.31	24	5.35E-03	−9.14E-04	5.62E-03	8.72E-01	M34040	20	14.53	19	7.52E-01	−6.82E-03	5.36E-03	2.20E-01
M35187	12	9.16	11	6.07E-01	8.53E-03	7.45E-03	2.79E-01	M35159	9	7.23	8	5.12E-01	6.34E-03	1.90E-02	7.49E-01
M35669	25	24.81	24	4.16E-01	5.72E-03	4.93E-03	2.58E-01	M37058	48	44.02	47	5.97E-01	3.99E-03	4.00E-03	3.23E-01
AVB								AP
M33415	25	11.95	24	9.81E-01	4.92E-03	5.58E-03	3.83E-01	M27738	18	21.11	17	2.21E-01	−7.75E-03	4.18E-03	8.24E-02
M37202	27	25.27	26	5.04E-01	−1.28E-03	1.04E-02	9.04E-01	M32691	16	11.52	15	7.15E-01	4.34E-03	8.61E-03	6.22E-01
PT								M32808	16	37.88	15	9.40E-04	−7.96E-03	9.15E-03	3.99E-01
M15140	44	41.92	43	5.18E-01	4.92E-03	5.58E-03	3.83E-01	M32846	25	26.23	24	3.42E-01	1.79E-03	4.05E-03	6.62E-01
M19364	20	18.87	19	4.65E-01	−1.28E-03	1.04E-02	9.04E-01	M32855	23	40.34	22	9.87E-03	−2.56E-03	6.91E-03	7.15E-01
M32388	7	7.03	6	3.18E-01	−5.87E-02	2.66E-02	7.85E-02	M32863	13	13.92	12	3.06E-01	−1.21E-02	6.76E-03	1.02E-01
M32855	23	22.22	22	4.47E-01	−1.03E-02	8.47E-03	2.36E-01	M32980	12	11.65	11	3.91E-01	9.10E-03	8.30E-03	2.99E-01
M34912	46	37.56	45	7.77E-01	−4.13E-03	5.53E-03	4.59E-01	M33230	14	11.17	13	5.96E-01	1.43E-02	9.60E-03	1.61E-01
AAA								M33447	11	8.09	10	6.20E-01	1.31E-02	1.01E-02	2.26E-01
M33973	16	7.50	15	9.42E-01	7.03E-04	1.18E-02	9.53E-01	M35160	10	9.88	9	3.60E-01	−8.81E-03	1.21E-02	4.88E-01
M35464	17	16.26	16	4.35E-01	−1.02E-02	2.50E-02	6.88E-01	M35187	12	15.18	11	1.74E-01	−9.25E-04	8.45E-03	9.15E-01
M37190	21	17.43	20	6.25E-01	−2.47E-03	1.16E-02	8.34E-01	M35669	25	34.57	24	7.50E-02	4.14E-03	5.44E-03	4.54E-01
MI								ICH
M27718	11	11.97	10	2.87E-01	−4.39E-03	1.05E-02	6.86E-01	M00584	29	27.05	28	5.16E-01	−1.61E-02	1.39E-02	2.57E-01
M32338	35	43.93	34	1.18E-01	1.58E-03	4.49E-03	7.27E-01	M31522	4	2.73	3	4.35E-01	−5.52E-02	4.63E-02	3.55E-01
M32808	16	32.88	15	4.87E-03	3.81E-04	1.03E-02	9.71E-01	M32560	12	9.43	11	5.82E-01	−6.56E-03	1.72E-02	7.10E-01
M32846	25	34.65	24	7.38E-02	9.20E-03	5.13E-03	8.58E-02	M32857	7	5.58	6	4.71E-01	1.97E-02	3.35E-02	5.82E-01
M32855	23	40.28	22	1.00E-02	−3.56E-04	8.13E-03	9.65E-01	M33627	10	7.73	9	5.62E-01	−3.69E-02	3.14E-02	2.74E-01
M33447	11	21.27	10	1.93E-02	2.87E-03	1.82E-02	8.78E-01	DDC
M33801	9	17.57	8	2.47E-02	−2.91E-02	1.60E-02	1.12E-01	M11438	4	2.31	3	5.11E-01	4.85E-03	1.45E-02	7.70E-01
M35160	10	9.34	9	4.06E-01	−4.96E-03	1.43E-02	7.37E-01	M32553	17	20.42	16	2.02E-01	2.35E-02	1.36E-02	1.04E-01
M35240	7	8.93	6	1.78E-01	1.34E-02	1.85E-02	5.00E-01	M32786	19	10.34	18	9.20E-01	1.73E-02	1.41E-02	2.36E-01
M35669	25	37.48	24	3.91E-02	4.48E-03	6.65E-03	5.08E-01	M32910	60	51.59	59	7.42E-01	4.52E-03	3.09E-03	1.49E-01
								M35854	13	9.67	12	6.45E-01	−2.26E-03	8.78E-03	8.02E-01

Abbreviations: nSNP, number of single nucleotide polymorphisms; IVW, inverse variance weighted; Q df, degrees of freedom for the Q; SE, standard error.

Metabolic Pathway Analysis

We did not find any significant pathways for known positive metabolites, as indicated by FDR and p-values exceeding .05. However, a notable exception was the pathway related to ‘vitamin B6 metabolism’, which was correlated with PE outcomes, exhibiting an actual P-value of .01733 (< .05). This finding shares some common ground with a previous randomised controlled trial (RCT),³⁵ emphasising that supplementing the vitamin B6 level could reduce homocysteine but not the risk of venous thrombosis recurrence. Additionally, an animal study³⁶ suggested that the vitamin B6 metabolite pyridoxal 5′-phosphate may alleviate intracellular Ca²⁺ overload and inhibit purinergic receptors, thereby reducing coronary artery occlusion and providing cardiac protection in ischaemic heart disease. Taken together, these studies underscore the broader cardiovascular impacts of vitamin B6.

Our enrichment analysis identified a significant association between the ‘glycine, serine and threonine metabolic’ pathway and MI outcomes, with an FDR of 0.0376 (< 0.05). This finding aligns with previous research that linked glycine levels with a lower risk of coronary heart disease (CHD), suggesting a protective role for glycine in CHD.³⁷ For an in-depth overview of the metabolites and pathways involved, please refer to Supplementary Tables S4–S6 and Supplementary Figures S5 and S6.

Discussion

We utilised large-scale public GWAS data and two-sample MR methods to explore the causal relations between blood metabolites and 10 CVDs. Beyond the initial primary IVW analysis, we also used four additional MR models along with sensitivity analyses to exclude confounding factors and enhance the reliability of our findings. This study may be the first to combine metabolomics and genomics to investigate the causal link between metabolites and CVDs. Our results identified causal relations between 36 known metabolites and various CVDs. Additionally, we discovered that the vitamin B6 metabolic pathway (involved metabolites: pyridoxate), as well as the glycine, serine and threonine metabolism set (involved metabolites: glycine and creatine), may play significant roles in the development and progression of PE and MI, respectively.

CVDs pose a significant public health challenge due to their high incidence and prevalence rates. Early detection, diagnosis and treatment of these conditions are of paramount importance in managing and mitigating their severe health impacts. Despite many studies on the pathogenesis of CVDs, the intricate and diverse nature of CVDs means that a complete understanding of their mechanisms remains elusive. This highlights the necessity for ongoing research in this area to enhance our comprehension and develop more effective prevention and treatment strategies, thereby reducing the overall health and societal burden.

Research on blood metabolites has intensified,³⁸ particularly in the context of CVDs, making it a prominent focus in the scientific community.¹⁴ Some clinical and animal studies have highlighted the critical role of blood metabolites in the development and progression of CVDs, revealing the systemic nature of these conditions and identifying essential metabolic pathways, such as gut microbial co-metabolism, branched-chain amino acids, glycerophospholipid and cholesterol metabolism and inflammatory processes.^39–42 Despite these advancements, the complex interplay between blood metabolites and CVDs continues to be an active and deepening area of investigation.

Our study identified three amino acids, four lipids and one nucleotide linked to AF. First, among the amino acids, 2-aminooctanoic acid and N-acetylornithine exhibited a negative correlation with AF, whereas isovalerylcarnitine was positively correlated with AF. 2-Aminooctanoic acid, an alpha-amino acid with an extended carbon chain, was involved in synthesising biologically active compounds,⁴³ indicating its potential role in AF. N-acetylornithine is significant in the urea cycle and polyamine biosynthesis and influences liver detoxification and nitrogen metabolism, potentially affecting endothelial function and blood flow.⁴⁴ In contrast, isovalerylcarnitine, vital in fatty acid metabolism and the transport of fatty acids into mitochondria for energy production, was detected at higher levels in patients with coronary artery disease (CAD),⁴⁵ a pattern that closely mirrors the AF findings. Four lipids, specifically 3-dehydrocarnitine, butyrylcarnitine and dodecanedioate, may increase the risk of AF, whereas tetradecanedioate could decrease it. 3-Dehydrocarnitine is a derivative of carnitine that plays a crucial role in lipid metabolism, specifically in the transport and oxidation of fatty acids within mitochondria; however, its association with CVDs remains understudied. Similarly, butyrylcarnitine is a form of acylcarnitine that is essential for transporting butyric acid to mitochondria for β-oxidation, a process vital for energy production in muscle and heart tissues. Intriguingly, an increase in the plasma level of butyrylcarnitine has been associated with a heightened risk of CVDs in patients with type 2 diabetes mellitus, as indicated by a previous cross-sectional study.⁴⁶ Dodecanedioate is a dicarboxylic acid with a 12-carbon chain that plays a significant role in fatty acid metabolism. Emerging studies have suggested a potential relationship between low-density lipoprotein (LDL) and cholesterol.⁴⁷ Finally, tetradecanedioate has been identified as a biomarker of transporter function and a product of ω-oxidation of fatty acids⁴⁸ that has been recently highlighted as a protective factor in obstructive sleep apnoea which agrees with our observations and calls for further exploration.⁴⁹ Inosine is a crucial component of purine metabolism that modulates gene translation and RNA editing by the post-transcriptional conversion of adenosine to inosine. This metabolite is also a coronary dilator because it relaxes coronary arteries and exhibits inotropic action.⁵⁰ However, research in the area of atrial fibrillation is still lacking.

We found only one lipid, 4-androsten-3beta, 17beta-diol disulfate 1, which increased the risk of AVB. This compound is an important steroid hormone in the androgen pathway. A previous prospective study revealed a positive correlation between this steroid hormone and increased mortality rates from cardiovascular disease.⁵¹

We identified two risk factors following our investigation of PT. The first was the amino acid kynurenine. Kynurenine metabolites are closely associated with inflammation, energy homeostasis, apoptosis and oxidative stress.⁵² Additionally, increased serum levels of kynurenine metabolites are associated with various cardiovascular diseases, including heart disease, atherosclerosis and endothelial dysfunction.⁵³ The second was lipid dodecanedioate. Dodecanedioate is a dodecane dicarboxylic acid that plays a role in fatty acid and organic acid metabolism, affecting energy balance and cellular signalling. The metabolic pathways of this compound are closely associated with LDL and cholesterol.⁴⁷

We found two lipids positively correlated with AAA, such as 5alpha-androstan-3beta, 17beta-diol disulfate and epiandrosterone sulfate. These two lipids are involved in androgen metabolism.⁵⁴ Sex differences are a risk factor for AAA, with males playing a dominant role.⁵⁵ Previous studies have reported that the androgen receptor promotes AAA development by regulating the expression of interleukin-1α and transforming growth factor-β1.⁵⁶ However, the underlying causes and mechanisms remain unclear. Nevertheless, this study, for the first time, discovered that genetically determined 5alpha-androstan-3beta, 17beta-diol disulfate and epiandrosterone sulfate may provide new diagnostic and therapeutic targets for AAA while also highlighting the need to be cautious of some adverse effects.

We identified two amino acids, three lipids and one peptide with strong correlations to MI. Within this framework, the two amino acids, creatine and glycine, were protective factors, each playing a significant role in cardiac health. Creatine plays a pivotal role in cardiac contraction and energy metabolism. Creatine levels decrease during MI due to a reduction in creatine transporter expression and the breakdown of phosphocreatine to prevent the depletion of adenosine triphosphate. This decline is associated with reduced contractile reserve in the myocardium.⁵⁷ Second, glycine is inversely correlated with MI risk. This association may be closely linked to its roles in antioxidant reactions, purine synthesis and collagen formation.^37,58 We also identified enriched metabolic pathways involving creatine and glycine. Glycine, serine and threonine metabolism is intricately linked with MI (Supplementary Table S6 and Figure S6). Recent animal studies have revealed that metoprolol induces cardioprotective effects in a murine model of acute myocardial ischaemia by modulating this specific metabolic pathway.⁵⁹ This result provides innovative therapeutic and diagnostic targets for the clinical management of MI. Long-chain acylcarnitine oleoylcarnitine has been identified. Previous nested case-control studies have reported that this metabolite exhibits the strongest correlation with cardiovascular mortality in dialysis patients, and this association persists even after multivariate adjustment with a direct link to cardiovascular death within 1 year.⁶⁰ Palmitoleic acid (16:1n7) is a regulator of physiological cardiac hypertrophy.^61,62 Furthermore, it may protect against cardiac fibrosis and inflammation.⁶³ Consistent with our previous discussion, tetradecanedioate is not only a protective factor against AF but also against MI. The final unique peptide, ADpSGEGDFXAEGGGVR, is a fibrinogen cleavage peptide that has been associated with osteoarthritis,⁶⁴ prostate cancer⁶⁵ and stroke.⁶⁶

We identified 1,5-AG and mannitol in PE. The former is a monosaccharide that is structurally similar to glucose and is a short-term biomarker for the metabolic control of diabetes.⁶⁷ In this study, 1,5-AG was positively correlated with the incidence of PE. Previous studies have also indicated its role as a risk factor for conditions such as osteoporosis⁶⁸ and dentures.⁶⁹ In contrast, mannitol is a naturally occurring sugar alcohol found in the body, primarily derived from artificial sweeteners.⁷⁰ Unfortunately, reports on the association between this carbohydrate and PE are sporadic. Recent studies have identified a positive correlation between mannitol and depression, indicating specific differences from our findings, thereby necessitating further in-depth research.⁷¹ Subsequently, our study identified pyridoxate, a decomposition product of vitamin B6, as a protective cofactor and vitamin in PE. Vitamin B6 is in circulation in the active form of pyridoxal-5′-phosphate.⁷² Recent studies have highlighted pyridoxate as a novel potential biomarker of pulmonary tuberculosis.^73,74 Furthermore, our metabolic pathway analysis suggests a possible association between pyridoxate, a participant in vitamin B6 metabolism, and PE (Supplementary Table S5 and Figure S5). As previously noted, a large-scale Dutch RCT reported that B-vitamin supplementation lowers homocysteine levels,⁷⁵ which are associated with an increased risk of deep vein thrombosis and PE. However, that trial did not establish a direct link between vitamin B6 and deep vein thrombosis or PE.³⁵ Similarly, another RCT reported that B-vitamin supplementation in patients with venous thromboembolism does not affect specific coagulation markers, fibrinolysis, or endothelial activation. However, increased levels of tissue plasminogen activator and plasminogen activator inhibitor-1 were observed in patients with higher baseline homocysteine levels.⁷⁵ Therefore, further research in this area remains crucial. Three lipids associated with PE were identified, including two risk factors: 1-arachidonoylglycerophosphocholine and arachidonate (20:4n6). 1-Arachidonoylglycerophosphocholine is functionally related to arachidonic acid, and its metabolic pathway is closely linked to the mechanism of the statin class of lipid-lowering drugs.⁷⁶ Arachidonate (20:4n6) is an unsaturated, essential fatty acid that is synthesised from dietary linoleic acid and serves as a precursor to prostaglandins, thromboxanes and leukotrienes. This metabolite and its derivatives play key roles in linking nutrient metabolism to immunity and inflammation and are crucial in the development and progression of CVDs.⁷⁷ Additionally, arachidonic acid and its monohydroperoxy- and hydroxy-metabolites exert inhibitory effects on procoagulant activity in endothelial cells, suggesting a role in modulating blood clotting and vital in the context of deep vein thrombosis and PE.⁷⁸ Contrary to two previous studies, the lysophospholipid 1-linoleoylglycerophosphoethanolamine reduces the risk of PE. Moreover, some studies have reported that this metabolite has a potential causal negative correlation with preeclampsia.⁷⁹ However, the specific mechanism of action remains unclear.

Four TIA protective factors were identified, including two amino acids, 2-aminooctanoic acid and cysteine-glutathione disulfide, and two energy-related metabolites, citrate and succinylcarnitine. Furthermore, the xenobiotic ibuprofen increases the risk of TIA. First, the α-amino fatty acid 2-aminooctanoic acid is a secondary metabolite that may function as a defence or signalling molecule. However, few reports have linked this metabolite with CVDs. Only a few studies have explored its involvement in the metabolic pathways linked to colorectal cancer.⁸⁰ Second, the peptide cysteine-glutathione disulfide has been linked in a recent case-cohort study from China to nonalcoholic fatty liver disease and was negatively correlated with BMI.⁸¹ Overweight and obesity are established risk factors for CVDs.^82,83 This finding is consistent with our conclusions and could represent the potential mechanism through which this metabolite contributes to the onset of TIAs. Citrate and succinylcarnitine are two energy-related metabolites. Citrate is a tricarboxylic acid and an intermediate in the central metabolic pathway of the tricarboxylic acid cycle. This metabolite may suppress inflammation^84,85 and it positively affects cardiovascular health.⁸⁵ These aspects bear similarities to the conclusions presented in this study. The latter, succinylcarnitine, is an O-acylcarnitine, and its association with improved glucose tolerance has been reported.^86–88 This could potentially contribute to its role as a protective factor against TIA. As a final point, the current study identified ibuprofen as a xenobiotic linked to increased TIA risk. Ibuprofen, a monocarboxylic acid widely used as a non-steroidal anti-inflammatory drug, functions as a cyclooxygenase inhibitor. Supporting our findings, observational studies have revealed that extensive use of this xenobiotic intensifies the risks of cardiovascular events and ischaemic stroke.^89,90

We identified six metabolites closely associated with AP. These metabolites include one cofactor or vitamin and five lipids. Threonate, the only metabolite in the cofactor and vitamin category, lowers the risk of AP. It is an ascorbate metabolite synthesised from the degradation of vitamin C.⁹¹ Additionally, threonate is indirectly linked and negatively correlated with hypertension,⁹² which may be a potential mechanism for its protective effect against AP. Among the remaining lipids, four were identified as protective factors. Notably, the lysophospholipid 1-palmitoleoylglycerophosphocholine displays various protective and anti-inflammatory effects. Higher levels of this compound promote the expression of cyclooxygenase-2 and endothelial nitric oxide synthase in endothelial cells, enhancing vascular protection via prostacyclin or nitric oxide production.⁹³ Furthermore, adrenate (22:4n6) is an oxylipin derived from arachidonic acid (20:4n-6) that is converted to epoxydocosatrienoic acid and mitigates endoplasmic reticulum stress and neuroinflammation.^94,95 Both lipids have been associated with a reduced risk of lacunar stroke, consistent with our findings.⁹⁶ Subsequently, the omega-7 monounsaturated fatty acid palmitoleate (16:1n7) and glycerides offer protective effects against cardiac fibrosis and inflammation, as previously reported.⁶³ Similarly, tetradecanedioate is classified as a long-chain fatty acid that has been frequently reported as a protective factor against various CVDs. Last, oleoylcarnitine is a risk factor for AP that aligns with other lipids associated with an increased risk of MI.

We identified mannose as a carbohydrate that reduces the risk of ICH. Mannose is a C2 epimer of D-glucose that is recognised as a natural bioactive monosaccharide and a principal glycoprotein involved in N-glycosylation reactions.⁹⁷ It inhibits oxidative bursts of neutrophils and plays a significant role in inflammation.⁹⁸ Furthermore, mannose is a protective factor in acute respiratory distress syndrome.⁹⁹ We also discovered pyroglutamylglycine as a peptide that increases ICH risk. It is an alpha-amino acid dipeptide composed of glycine and 5-oxo-L-proline, linked with a peptide bond. A previous prospective study revealed a correlation between pyroglutamylglycine and pancreatic ductal adenocarcinoma, identifying it as a risk factor.¹⁰⁰

In the final analysis, we identified four distinct categories of blood metabolites significantly associated with DCC. Phenol sulfate is an aryl sulfate with an O-sulfo substituent on phenol and is classified as an amino acid. It is a uremic toxin¹⁰¹ that acts as a protein-bound uremic solute. This compound stimulates the production of reactive oxygen species and reduces glutathione levels, thus increasing cellular susceptibility to oxidative stress.¹⁰² This may be an underlying reason for its identification as a risk factor for DCC. In contrast, the remaining three metabolites were identified as protective factors against DCC. The carbohydrate threitol is the main endproduct of D-xylose metabolism. Changes in threitol and other polyols occur in diseases such as diabetes mellitus, which involves disruptions in carbohydrate metabolism.¹⁰³ Moreover, O-methylascorbate was a recognised metabolic derivative of L-ascorbic acid (vitamin C).¹⁰⁴ Previous studies indicated that vitamin C levels are inversely correlated with systemic and pulmonary blood pressure.^105–107 Phosphate is vital in numerous physiological and pathological processes, including energy metabolism, cellular structural integrity and signal transduction pathways.^108–110 Previous research has thoroughly investigated the epidemiological links between phosphate levels and diverse cardiovascular events.^111–114 Nevertheless, the conclusions continue to be inconsistent.

Across CVD outcomes, accumulating metabolomics evidence suggests several converging metabolic themes that may reflect shared biological mechanisms.^39,115 In particular, dysregulated amino acid metabolism — including pathways involving glycine, branched-chain amino acids (BCAAs), aromatic amino acids and their derivatives, as well as kynurenine-pathway metabolites — has been repeatedly associated with atherosclerotic cardiovascular disease (ASCVD), implicating disturbances in nitrogen balance, inflammation, endothelial function and lipid metabolism.^116,117 Meanwhile, perturbations in lipid-related metabolism — notably alterations in acylcarnitines and fatty-acid oxidation intermediates — point to deranged mitochondrial β-oxidation, impaired fatty acid transport, and disrupted lipid-mediated signalling as common features in CVD-related pathophysiology.¹¹⁸ Additional evidence supports roles for metabolites tied to oxidative stress, endothelial dysfunction and pro-inflammatory lipid mediators.¹¹⁹ Although the precise metabolite signatures vary by CVD subtype, these findings support a unifying metabolic network underpinning cardiometabolic homeostasis and disease susceptibility — a network that may offer novel biomarkers or therapeutic targets across multiple CVD phenotypes.^120,121

Our study had several strengths. This inaugural work explored the potential causal relations between blood metabolites as exposure factors and the risks of diverse CVDs employing MR. This approach substantially mitigated the biases inherent in traditional observational studies, offering significant clinical research value. Additionally, our use of extensive and diverse GWAS databases that were filtered to yield variables with an F- statistic of >10 effectively reduced the tool bias and enhanced the interpretive power of our research. This study further employed stringent quality control conditions, multiple models for evaluating causal effects and comprehensive, detailed sensitivity analyses, ensuring reliable results. Third, unlike previous single-exposure MR analyses, analysing a multitude of blood metabolites posed a substantial workload and analytical challenges. Our proposed strategy offers a reference for similar research endeavours. Nevertheless, this study also faced some limitations. First, our GWAS data were derived exclusively from European ancestry, so caution is warranted when generalising these findings to other ethnic groups. Second, the uniqueness of the blood metabolite data precluded a precise calculation of sample overlap in our study, potentially affecting the results. Third, due to limitations in the KEGG and SMPDB databases, this study focused on the causal relations of known downstream KEGG and SMPDB metabolites with CVDs and did not exclude the effects via currently unknown or unincorporated KEGG pathways and SMPDB. Fourth, a more lenient clumping criterion was adopted, given the smaller sample size of the exposure data. Finally, due to constraints in the available outcome data, we were unable to include all cardio-cerebrovascular phenotypes or perform more detailed subtype analyses. We hope that future research will benefit from larger sample sizes and more refined outcome definitions to further extend and validate our findings. In conclusion, while our results are significant, these limitations must be considered when interpreting outcomes and determining their broader applicability.

To address these limitations and further elucidate the complex relations between blood metabolites and CVD, scholars should utilise more varied metabolite data, more comprehensive and in-depth analytical approaches, and precise experimental methodologies in future validations.

In summary, our study demonstrated a strong association between various metabolites and different CVDs, offering pathways for future research to improve the diagnosis, prevention and treatment of CVDs. However,studies linking these specific metabolites to various cardiovascular diseases are still relatively scarce, and the proposed mechanisms are largely speculative. Therefore, more in-depth research is essential.

Conclusion

This study is the first to systematically apply MR using large-scale genomic data to investigate the causal relationships between blood metabolites and multiple CVD phenotypes. Our analyses provide preliminary evidence that circulatory metabolic disturbances may contribute to CVD risk. Through multiple MR methods and extensive sensitivity testing, we identified significant associations between 36 metabolites and 10 CVD phenotypes. Pathway enrichment analyses further highlighted two metabolic pathways that may play key roles in PE and MI. Overall, these findings offer new insights into the metabolic underpinnings of CVDs and suggest potential biomarkers and therapeutic targets for future research.

Supplemental Material

sj-xlsx-1-cat-10.1177_10760296261420227 - Supplemental material for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study

Supplemental material, sj-xlsx-1-cat-10.1177_10760296261420227 for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study by Honghong Zhang, Jiangzhen Xie, Chaojie He, Huilin Hu, Changlin Zhai, Gang Qian and Menghui Mao in Clinical and Applied Thrombosis/Hemostasis

Supplemental Material

sj-pdf-2-cat-10.1177_10760296261420227 - Supplemental material for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study

Supplemental material, sj-pdf-2-cat-10.1177_10760296261420227 for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study by Honghong Zhang, Jiangzhen Xie, Chaojie He, Huilin Hu, Changlin Zhai, Gang Qian and Menghui Mao in Clinical and Applied Thrombosis/Hemostasis

Supplemental Material

sj-pdf-3-cat-10.1177_10760296261420227 - Supplemental material for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study

Supplemental material, sj-pdf-3-cat-10.1177_10760296261420227 for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study by Honghong Zhang, Jiangzhen Xie, Chaojie He, Huilin Hu, Changlin Zhai, Gang Qian and Menghui Mao in Clinical and Applied Thrombosis/Hemostasis

Supplemental Material

sj-pdf-4-cat-10.1177_10760296261420227 - Supplemental material for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study

Supplemental material, sj-pdf-4-cat-10.1177_10760296261420227 for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study by Honghong Zhang, Jiangzhen Xie, Chaojie He, Huilin Hu, Changlin Zhai, Gang Qian and Menghui Mao in Clinical and Applied Thrombosis/Hemostasis

Supplemental Material

sj-pdf-5-cat-10.1177_10760296261420227 - Supplemental material for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study

Supplemental material, sj-pdf-5-cat-10.1177_10760296261420227 for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study by Honghong Zhang, Jiangzhen Xie, Chaojie He, Huilin Hu, Changlin Zhai, Gang Qian and Menghui Mao in Clinical and Applied Thrombosis/Hemostasis

Supplemental Material

sj-pdf-6-cat-10.1177_10760296261420227 - Supplemental material for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study

Supplemental material, sj-pdf-6-cat-10.1177_10760296261420227 for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study by Honghong Zhang, Jiangzhen Xie, Chaojie He, Huilin Hu, Changlin Zhai, Gang Qian and Menghui Mao in Clinical and Applied Thrombosis/Hemostasis

Supplemental Material

sj-pdf-7-cat-10.1177_10760296261420227 - Supplemental material for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study

Supplemental material, sj-pdf-7-cat-10.1177_10760296261420227 for Causal Associations Between Human Blood Metabolites and Cardio-Cerebrovascular Diseases: A Mendelian Randomisation Study by Honghong Zhang, Jiangzhen Xie, Chaojie He, Huilin Hu, Changlin Zhai, Gang Qian and Menghui Mao in Clinical and Applied Thrombosis/Hemostasis

Footnotes

Acknowledgements

We extend our appreciation to the participants and researchers involved in the Metabolomics GWAS Server and the FinnGen consortium.We thank all participants and investigators for sharing these data.

ORCID iD

Menghui Mao

Ethics Approval and Consent to Participate

Not applicable to this study. Written informed consent and approval from the local ethical committee were obtained by all included GWAS studies.

Author Contributions

HZ, JX, and MM designed the study, HZ, JX and CH analyzed the data and drafted the article. CZ, HH and MM edited the manuscript. GQ and MM supervised the study and acquired funding for the work. All authors read and approved the published version of the manuscript.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by the Zhejiang Provincial Natural Science Foundation of China under Grant No. LQ23H020001, National Natural Science Foundation of China (No. 82300363), Zhejiang Provincial Natural Science Foundation of China under Grant No. LGF21H020006, 2023 Zhe-jiang Province Traditional Chinese Medicine Scientific Research Fund (2023ZL700), The 2019 key medical disciplnes jointly established by provinces and city, China (2019-ss-xxgbx), Jiaxing Key Laboratory of Arteriosclerotic Diseases (2020-dmzdsys), Pioneer innovation team of Jiaxing Institute of Arteriosclerotic Disease (XFCX-DMYH), Jiaxing First Hospital independent research project (ZZKT2022-001, ZZKT2022-003, ZZKT2022-004, ZZKT2022-005, ZZKT2022-007, ZZKT2022-008).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The data that support the findings of this study are openly available in FinnGen and Metabolomics at FinnGen Results and mGWAS.

Supplemental Material

Supplemental material for this article is available online.

References

Organization WH. Cardiovascular diseases (CVDs) [EB/OL], https://www.who.int/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds)#:∼:text=Cardiovascular%20diseases%20,17%20million%20premature%20deaths (11 June 2021).

Jan. EOLBDBA. Cardiovascular Disease, https://www.ncbi.nlm.nih.gov/books/NBK535419/.

Kapuku

Kop

. Classification of cardiovascular diseases: epidemiology, diagnosis, and treatment. In: Waldstein

Kop

Suarez

, eds. Handbook of cardiovascular behavioral medicine. Springer New York; 2022:45–80.

Fuster

Kelly

, eds. Promoting cardiovascular health in the developing world: a critical challenge to achieve global health. National Academies Press; 2010.

Twig

Yaniv

Levine

, et al. Body-mass index in 2.3 million adolescents and cardiovascular death in adulthood. N Engl J Med. 2016;374(25):2430–2440. doi:10.1056/NEJMoa1503840

Jensen

Ryan

Apovian

, et al. 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: A report of the American College of Cardiology/American Heart Association task force on practice guidelines and the obesity society. Circulation. 2014;129(25 Suppl 2):S102–S138. doi:10.1161/01.cir.0000437739.71477.ee

Al-Delaimy

Merchant

Rimm

, et al. Effect of type 2 diabetes and its duration on the risk of peripheral arterial disease among men. Am J Med. 2004;116(4):236–240. doi:10.1016/j.amjmed.2003.09.038

Gerstein

Pais

Pogue

, et al. Relationship of glucose and insulin levels to the risk of myocardial infarction: A case-control study. J Am Coll Cardiol. 1999;33(3):612–619. doi:10.1016/s0735-1097(98)00637-8

Singer

Nathan

Anderson

, et al. Association of HbA1c with prevalent cardiovascular disease in the original cohort of the framingham heart study. Diabetes. 1992;41(2):202–208. doi:10.2337/diab.41.2.202

10.

Yusuf

Hawken

Ounpuu

, et al. Effect of potentially modifiable risk factors associated with myocardial infarction in 52 countries (the INTERHEART study): Case-control study. Lancet. 2004;364(9438):937–952. doi:10.1016/S0140-6736(04)17018-9

11.

Prescott

Hippe

Schnohr

, et al. Smoking and risk of myocardial infarction in women and men: Longitudinal population study. BMJ. 1998;316(7137):1043–1047. doi:10.1136/bmj.316.7137.1043

12.

Takeda

Aotani

Hashimoto

, et al. Association of heavy smoking with high body mass index and cardiovascular risk score in Japan. Obes Sci Pract. 2025;11(4):e70089. doi:10.1002/osp4.70089

13.

Chen

. Guide to metabolomics analysis: A bioinformatics workflow. Metabolites. 2022;12(4):357. doi:10.3390/metabo12040357

14.

McGranaghan

Kirwan

Garcia-Rivera

, et al. Lipid metabolite biomarkers in cardiovascular disease: Discovery and biomechanism translation from human studies. Metabolites. 2021;11(9):621. doi:10.3390/metabo11090621

15.

Cheng

Liu

Sun

, et al. Integrative metabolomics and genomics reveal molecular signatures for type 2 diabetes and its cardiovascular complications. Cardiovasc Diabetol. 2025;24(1):166. doi:10.1186/s12933-025-02718-4

16.

Islam

Liu

Mohandas

, et al. Metabolomic signatures of ideal cardiovascular health in black adults. Sci Rep. 2024;14(1):1794. doi:10.1038/s41598-024-51920-z

17.

Shin

Fauman

Petersen

, et al. An atlas of genetic influences on human blood metabolites. Nat Genet. 2014;46(6):543–550. doi:10.1038/ng.2982

18.

Priori

Aliot

Blomstrom-Lundqvist

, et al. Update of the guidelines on sudden cardiac death of the European Society of Cardiology. Eur Heart J. 2003;24(1):13–15.

19.

Maitre

Shemie

Baker

, et al. Knowledge gaps in the definition and determination of death. Can J Anaesth. 2023;70(4):610–616. doi:10.1007/s12630-023-02422-5

20.

Kurki

Karjalainen

Palta

, et al. Finngen provides genetic insights from a well-phenotyped isolated population. Nature. 2023;613(7944):508–518. doi:10.1038/s41586-022-05473-8

21.

J-H

Yang

Y-H

Wang

Y-C

, et al. Causal effects of blood metabolites and obstructive sleep apnea: A Mendelian randomization study. Curr Neurovasc Res. 2024;21(1):101–109. doi:10.2174/0115672026266627230921052416

22.

Papadimitriou

Dimou

Tsilidis

, et al. Physical activity and risks of breast and colorectal cancer: A Mendelian randomisation analysis. Nat Commun. 2020;11(1):597. doi:10.1038/s41467-020-14389-8

23.

Bowden

Del Greco

Minelli

, et al. A framework for the investigation of pleiotropy in two-sample summary data Mendelian randomization. Stat Med. 2017;36(11):1783–1802. doi:10.1002/sim.7221

24.

Bowden

Spiller

Del Greco

, et al. Improving the visualization, interpretation and analysis of two-sample summary data Mendelian randomization via the radial plot and radial regression. Int J Epidemiol. 2018;47(4):1264–1278. doi:10.1093/ije/dyy101

25.

Bowden

Del Greco

Minelli

, et al. Assessing the suitability of summary data for two-sample Mendelian randomization analyses using MR-egger regression: The role of the I2 statistic. Int J Epidemiol. 2016;45(6):1961–1974. doi:10.1093/ije/dyw220

26.

Bowden

Davey Smith

Haycock

, et al. Consistent estimation in Mendelian randomization with some invalid instruments using a weighted median estimator. Genet Epidemiol. 2016;40(4):304–314. doi:10.1002/gepi.21965

27.

Milne

Kuchenbaecker

Michailidou

, et al. Identification of ten variants associated with risk of estrogen-receptor-negative breast cancer. Nat Genet. 2017;49(12):1767–1778. doi:10.1038/ng.3785

28.

Greco

Minelli

Sheehan

, et al. Detecting pleiotropy in Mendelian randomisation studies with summary data and a continuous outcome. Stat Med. 2015;34(21):2926–2940. doi:10.1002/sim.6522

29.

Burgess

Thompson

. Interpreting findings from Mendelian randomization using the MR-egger method. Eur J Epidemiol. 2017;32(5):377–389. doi:10.1007/s10654-017-0255-x

30.

Zheng

Baird

Borges

, et al. Recent developments in Mendelian randomization studies. Curr Epidemiol Rep. 2017;4(4):330–345. doi:10.1007/s40471-017-0128-6

31.

Davey Smith

Hemani

. Mendelian randomization: Genetic anchors for causal inference in epidemiological studies. Hum Mol Genet. 2014;23(R1):R89–R98. doi:10.1093/hmg/ddu328

32.

Verbanck

Chen

Neale

, et al. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat Genet. 2018;50(5):693–698. doi:10.1038/s41588-018-0099-7

33.

Frolkis

Knox

Lim

, et al. SMPDB: The small molecule pathway database. Nucleic Acids Res. 2010;38:D480–D487. doi:10.1093/nar/gkp1002

34.

Kanehisa

Goto

Sato

, et al. KEGG For integration and interpretation of large-scale molecular data sets. Nucleic Acids Res. 2012;40:D109–D114. doi:10.1093/nar/gkr988

35.

den Heijer

Willems

Blom

, et al. Homocysteine lowering by B vitamins and the secondary prevention of deep vein thrombosis and pulmonary embolism: A randomized, placebo-controlled, double-blind trial. Blood. 2007;109(1):139–144. doi:10.1182/blood-2006-04-014654

36.

Dhalla

Takeda

Elimban

. Mechanisms of the beneficial effects of vitamin B6 and pyridoxal 5-phosphate on cardiac performance in ischemic heart disease. Clin Chem Lab Med. 2013;51(3):535–543. doi:10.1515/cclm-2012-0553

37.

Wittemans

LBL

Lotta

Oliver-Williams

, et al. Assessing the causal association of glycine with risk of cardio-metabolic diseases. Nat Commun. 2019;10(1):1060. doi:10.1038/s41467-019-08936-1

38.

Buergel

Steinfeldt

Ruyoga

, et al. Metabolomic profiles predict individual multidisease outcomes. Nat Med. 2022;28(11):2309–2320. doi:10.1038/s41591-022-01980-3

39.

Iliou

Mikros

Karaman

, et al. Metabolic phenotyping and cardiovascular disease: An overview of evidence from epidemiological settings. Heart. 2021;107(14):1123–1129. doi:10.1136/heartjnl-2019-315615

40.

Bugger

Byrne

Abel

. Animal models of dysregulated cardiac metabolism. Circ Res. 2022;130(12):1965–1993. doi:10.1161/CIRCRESAHA.122.320334

41.

McGarrah

Crown

Zhang

, et al. Cardiovascular metabolomics. Circ Res. 2018;122(9):1238–1258. doi:10.1161/CIRCRESAHA.117.311002

42.

Shah

Kraus

Newgard

. Metabolomic profiling for the identification of novel biomarkers and mechanisms related to common cardiovascular diseases: Form and function. Circulation. 2012;126(9):1110–1120. doi:10.1161/CIRCULATIONAHA.111.060368

43.

Almahboub

Narancic

Devocelle

, et al. Biosynthesis of 2-aminooctanoic acid and its use to terminally modify a lactoferricin B peptide derivative for improved antimicrobial activity. Appl Microbiol Biotechnol. 2018;102(2):789–799. doi:10.1007/s00253-017-8655-0

44.

Oratz

Rothschild

Schreiber

, et al. The role of the urea cycle and polyamines in albumin synthesis. Hepatology. 1983;3(4):567–571. doi:10.1002/hep.1840030415

45.

Gander

Carrard

Gallart-Ayala

, et al. Metabolic impairment in coronary artery disease: Elevated Serum acylcarnitines under the spotlights. Front Cardiovasc Med. 2021;8:792350. doi:10.3389/fcvm.2021.792350

46.

Zhao

Feng

Huang

, et al. The association between acylcarnitine metabolites and cardiovascular disease in Chinese patients with type 2 diabetes Mellitus. Front Endocrinol (Lausanne). 2020;11:212. doi:10.3389/fendo.2020.00212

47.

Fernandes Silva

Ravi

Vangipurapu

, et al. Effects of SLCO1B1 genetic variant on metabolite profile in participants on simvastatin treatment. Metabolites. 2022;12(12):1159. doi:10.3390/metabo12121159

48.

Sun

Tiedt

, et al. A prospective study of serum metabolites and risk of ischemic stroke. Neurology. 2019;92(16):e1890–e1898. doi:10.1212/WNL.0000000000007279

49.

J-H

Yang

Y-H

Wang

Y-C

, et al. Causal effects of blood metabolites and obstructive sleep apnea: A Mendelian randomization study. Curr Neurovasc Res. 2024;21(1):101–109. doi:10.2174/0115672026266627230921052416

50.

Kim

. Inosine: A bioactive metabolite with multimodal actions in human diseases. Front Pharmacol. 2022;13:1043970. doi:10.3389/fphar.2022.1043970

51.

Huang

Weinstein

Moore

, et al. Serum metabolomic profiling of all-cause mortality: A prospective analysis in the alpha-tocopherol, beta-carotene cancer prevention (ATBC) study cohort. Am J Epidemiol. 2018;187(8):1721–1732. doi:10.1093/aje/kwy017

52.

Lund

Nordrehaug

Slettom

, et al. Plasma kynurenines and prognosis in patients with heart failure. PLoS One. 2020;15(1):e0227365. doi:10.1371/journal.pone.0227365

53.

Song

Ramprasath

Wang

, et al. Abnormal kynurenine pathway of tryptophan catabolism in cardiovascular diseases. Cell Mol Life Sci. 2017;74(16):2899–2916. doi:10.1007/s00018-017-2504-2

54.

Vehof

Hysi

Hammond

. A metabolome-wide study of dry eye disease reveals Serum androgens as biomarkers. Ophthalmology. 2017;124(4):505–511. doi:10.1016/j.ophtha.2016.12.011

55.

Hellmann

Grand

Freischlag

. Inflammatory abdominal aortic aneurysm. JAMA. 2007;297(4):395–400. doi:10.1001/jama.297.4.395

56.

Huang

Luo

Lai

, et al. Androgen receptor promotes abdominal aortic aneurysm development via modulating inflammatory interleukin-1alpha and transforming growth factor-beta1 expression. Hypertension. 2015;66(4):881–891. doi:10.1161/HYPERTENSIONAHA.115.05654

57.

Balestrino

. Role of creatine in the heart: Health and disease. Nutrients. 2021;13(4):1215. doi:10.3390/nu13041215

58.

Ding

Svingen

Pedersen

, et al. Plasma glycine and risk of acute myocardial infarction in patients with suspected stable angina pectoris. J Am Heart Assoc. 2015;5(1):e002621. doi:10.1161/JAHA.115.002621

59.

Lai

Yuan

Wang

, et al. Metabolomic profiling of metoprolol-induced cardioprotection in a murine model of acute myocardial ischemia. Biomed Pharmacother. 2020;124:109820. doi:10.1016/j.biopha.2020.109820

60.

Kalim

Clish

Wenger

, et al. A plasma long-chain acylcarnitine predicts cardiovascular mortality in incident dialysis patients. J Am Heart Assoc. 2013;2(6):e000542. doi:10.1161/JAHA.113.000542

61.

Riquelme

Magida

Harrison

, et al. Fatty acids identified in the Burmese python promote beneficial cardiac growth. Science. 2011;334(6055):528–531. doi:10.1126/science.1210558

62.

Foryst-Ludwig

Kreissl

Benz

, et al. Adipose tissue lipolysis promotes exercise-induced cardiac hypertrophy involving the lipokine C16:1n7-palmitoleate. J Biol Chem. 2015;290(39):23603–23615. doi:10.1074/jbc.M115.645341

63.

Betz

Qaiyumi

Goeritzer

, et al. Cardioprotective effects of palmitoleic acid (C16:1n7) in a mouse model of catecholamine-induced cardiac damage are mediated by PPAR activation. Int J Mol Sci. 2021;22(23):12695. doi:10.3390/ijms222312695

64.

Jin

, et al. Causality of genetically determined metabolites and metabolic pathways on osteoarthritis: A two-sample Mendelian randomization study. J Transl Med. 2023;21(1):357. doi:10.1186/s12967-023-04165-9

65.

Huang

Zhao

Weinstein

, et al. Metabolomic profile of prostate cancer-specific survival among 1812 Finnish men. BMC Med. 2022;20(1):362. doi:10.1186/s12916-022-02561-4

66.

Zhang

Cao

Zhao

, et al. Assessing the causal effect of genetically predicted metabolites and metabolic pathways on stroke. J Transl Med. 2023;21(1):822. doi:10.1186/s12967-023-04677-4

67.

Skupien

Gorczynska-Kosiorz

Klupa

, et al. Clinical application of 1,5-anhydroglucitol measurements in patients with hepatocyte nuclear factor-1alpha maturity-onset diabetes of the young. Diabetes Care. 2008;31(8):1496–1501. doi:10.2337/dc07-2334

68.

Liu

Wen

Zhang

, et al. Assessing the associations of blood metabolites with osteoporosis: A Mendelian randomization study. J Clin Endocrinol Metab. 2018;103(5):1850–1855. doi:10.1210/jc.2017-01719

69.

Zheng

Zhang

, et al. Circulating metabolites and dental traits: A Mendelian randomization study. J Dent Res. 2023;102(13):1460–1467. doi:10.1177/00220345231196536

70.

Zumbe

Lee

Storey

. Polyols in confectionery: The route to sugar-free, reduced sugar and reduced calorie confectionery. Br J Nutr. 2001;85(Suppl 1):S31–S45. doi:10.1079/bjn2000260

71.

van der Spek

Stewart

Kuhnel

, et al. Circulating metabolites modulated by diet are associated with depression. Mol Psychiatry. 2023;28(9):3874–3887. doi:10.1038/s41380-023-02180-2

72.

Zhang

Tsuchiya

Kinoshita

, et al. Vitamin B6 prevents IL-1beta protein production by inhibiting NLRP3 inflammasome activation. J Biol Chem. 2016;291(47):24517–24527. doi:10.1074/jbc.M116.743815

73.

Huang

Shi

Wei

, et al. Plasma metabolites Xanthine, 4-pyridoxate, and d-glutamic acid as novel potential biomarkers for pulmonary tuberculosis. Clin Chim Acta. 2019;498:135–142. doi:10.1016/j.cca.2019.08.017

74.

Dutta

Tornheim

Fukutani

, et al. Integration of metabolomics and transcriptomics reveals novel biomarkers in the blood for tuberculosis diagnosis in children. Sci Rep. 2020;10(1):19527. doi:10.1038/s41598-020-75513-8

75.

Rodrigues

Morelli

DAS

, et al. Homocysteine reduction by B-vitamin supplementation increases t-PA and PAI-1 levels in patients with venous thromboembolism. J Thromb Haemost. 2007;5(1):195–198. doi:10.1111/j.1538-7836.2006.02269.x

76.

Altmaier

Fobo

Heier

, et al. Metabolomics approach reveals effects of antihypertensives and lipid-lowering drugs on the human metabolism. Eur J Epidemiol. 2014;29(5):325–336. doi:10.1007/s10654-014-9910-7

77.

Sonnweber

Pizzini

Nairz

, et al. Arachidonic acid metabolites in cardiovascular and metabolic diseases. Int J Mol Sci. 2018;19(11):3285. doi:10.3390/ijms19113285

78.

Bates

Ferrante

Poulos

, et al. Inhibitory effects of arachidonic acid (20:4,n-6) and its monohydroperoxy- and hydroxy-metabolites on procoagulant activity in endothelial cells. Atherosclerosis. 1995;116(1):125–133. doi:10.1016/0021-9150(95)05538-8

79.

Ross

Baer

Ryckman

, et al. Second trimester inflammatory and metabolic markers in women delivering preterm with and without preeclampsia. J Perinatol. 2019;39(2):314–320. doi:10.1038/s41372-018-0275-8

80.

Brown

Rao

Weir

, et al. Metabolomics and metabolic pathway networks from human colorectal cancers, adjacent mucosa, and stool. Cancer Metab. 2016;4:11. doi:10.1186/s40170-016-0151-y

81.

Pang

Kartsonaki

, et al. Adiposity, metabolomic biomarkers, and risk of nonalcoholic fatty liver disease: A case-cohort study. Am J Clin Nutr. 2022;115(3):799–810. doi:10.1093/ajcn/nqab392

82.

Doehner

Schenkel

Anker

, et al. Overweight and obesity are associated with improved survival, functional outcome, and stroke recurrence after acute stroke or transient ischaemic attack: Observations from the TEMPiS trial. Eur Heart J. 2013;34(4):268–277. doi:10.1093/eurheartj/ehs340

83.

Bibbins-Domingo

Coxson

Pletcher

, et al. Adolescent overweight and future adult coronary heart disease. N Engl J Med. 2007;357(23):2371–2379. doi:10.1056/NEJMsa073166

84.

Oudemans-van Straaten

Bosman

Koopmans

, et al. Citrate anticoagulation for continuous venovenous hemofiltration. Crit Care Med. 2009;37(2):545–552. doi:10.1097/CCM.0b013e3181953c5e

85.

Tiranathanagul

Jearnsujitwimol

Susantitaphong

, et al. Regional citrate anticoagulation reduces polymorphonuclear cell degranulation in critically ill patients treated with continuous venovenous hemofiltration. Ther Apher Dial. 2011;15(6):556–564. doi:10.1111/j.1744-9987.2011.00996.x

86.

Noland

Koves

Seiler

, et al. Carnitine insufficiency caused by aging and overnutrition compromises mitochondrial performance and metabolic control. J Biol Chem. 2009;284(34):22840–22852. doi:10.1074/jbc.M109.032888

87.

Muoio

Noland

Kovalik

, et al. Muscle-specific deletion of carnitine acetyltransferase compromises glucose tolerance and metabolic flexibility. Cell Metab. 2012;15(5):764–777. doi:10.1016/j.cmet.2012.04.005

88.

Huffman

Koves

Hubal

, et al. Metabolite signatures of exercise training in human skeletal muscle relate to mitochondrial remodelling and cardiometabolic fitness. Diabetologia. 2014;57(11):2282–2295. doi:10.1007/s00125-014-3343-4

89.

Barcella

Lamberts

McGettigan

, et al. Differences in cardiovascular safety with non-steroidal anti-inflammatory drug therapy-A nationwide study in patients with osteoarthritis. Basic Clin Pharmacol Toxicol. 2019;124(5):629–641. doi:10.1111/bcpt.13182

90.

Schink

Kollhorst

Varas Lorenzo

, et al. Risk of ischemic stroke and the use of individual non-steroidal anti-inflammatory drugs: A multi-country European database study within the SOS project. PLoS One. 2018;13(9):e0203362. doi:10.1371/journal.pone.0203362

91.

Wang

Jiang

. Determination of L-threonate in human plasma and urine by high performance liquid chromatography-tandem mass spectrometry. J Chromatogr B Analyt Technol Biomed Life Sci. 2006;834(1–2):155–162. doi:10.1016/j.jchromb.2006.02.057

92.

Shi

, et al. Metabolomics study of blood pressure salt-sensitivity and hypertension. Nutr Metab Cardiovasc Dis. 2022;32(7):1681–1692. doi:10.1016/j.numecd.2022.04.002

93.

Knuplez

Marsche

. An updated review of pro- and anti-inflammatory properties of plasma lysophosphatidylcholines in the vascular system. Int J Mol Sci. 2020;21(12):4501. doi:10.3390/ijms21124501

94.

Dyall

Balas

Bazan

, et al. Polyunsaturated fatty acids and fatty acid-derived lipid mediators: Recent advances in the understanding of their biosynthesis, structures, and functions. Prog Lipid Res. 2022;86:101165. doi:10.1016/j.plipres.2022.101165

95.

Singh

Barnych

Wagner

, et al. Adrenic acid-derived epoxy fatty acids are naturally occurring lipids and their methyl ester prodrug reduces endoplasmic Reticulum stress and inflammatory pain. ACS Omega. 2021;6(10):7165–7174. doi:10.1021/acsomega.1c00241

96.

Guo

Hao

Tian

, et al. Human blood metabolites and lacunar stroke: A Mendelian randomization study. Int J Stroke. 2023;18(1):109–116. doi:10.1177/17474930221140792

97.

Alton

Hasilik

Niehues

, et al. Direct utilization of mannose for mammalian glycoprotein biosynthesis. Glycobiology. 1998;8(3):285–295. doi:10.1093/glycob/8.3.285

98.

Xie

Shen

, et al. Mannose prevents lipopolysaccharide-induced acute lung injury in rats. Inflamm Res. 2008;57(3):104–110. doi:10.1007/s00011-007-7037-y

99.

Wei

Huang

Zhang

, et al. Association of Serum mannose with acute respiratory distress syndrome risk and survival. JAMA Netw Open. 2021;4(1):e2034569. doi:10.1001/jamanetworkopen.2020.34569

100.

Stolzenberg-Solomon

Derkach

Moore

, et al. Associations between metabolites and pancreatic cancer risk in a large prospective epidemiological study. Gut. 2020;69(11):2008–2015. doi:10.1136/gutjnl-2019-319811

101.

Kanemitsu

Tsukamoto

Matsumoto

, et al. Generation and characterization of anti-phenyl sulfate monoclonal antibodies and a potential use for phenyl sulfate analysis in human blood. Biol Pharm Bull. 2018;41(8):1170–1177. doi:10.1248/bpb.b17-00925

102.

Edamatsu

Fujieda

Itoh

. Phenyl sulfate, indoxyl sulfate and p-cresyl sulfate decrease glutathione level to render cells vulnerable to oxidative stress in renal tubular cells. PLoS One. 2018;13(2):e0193342. doi:10.1371/journal.pone.0193342

103.

Aloia

. Monosaccharides and polyols in diabetes mellitus and uremia. J Lab Clin Med. 1973;82(5):809–817.

104.

Krumsiek

Suhre

Evans

, et al. Mining the unknown: A systems approach to metabolite identification combining genetic and metabolic information. PLoS Genet. 2012;8(10):e1003005. doi:10.1371/journal.pgen.1003005

105.

Rodrigo

Prat

Passalacqua

, et al. Relationship between oxidative stress and essential hypertension. Hypertens Res. 2007;30(12):1159–1167. doi:10.1291/hypres.30.1159

106.

Duvall

Pikman

Kantor

, et al. Pulmonary hypertension associated with scurvy and vitamin deficiencies in an autistic child. Pediatrics. 2013;132(6):e1699–e1703. doi:10.1542/peds.2012-3054

107.

Kupari

Rapola

. Reversible pulmonary hypertension associated with vitamin C deficiency. Chest. 2012;142(1):225–227. doi:10.1378/chest.11-1857

108.

Bhadada

Rao

. Role of phosphate in biomineralization. Calcif Tissue Int. 2021;108(1):32–40. doi:10.1007/s00223-020-00729-9

109.

Michigami

Kawai

Yamazaki

, et al. Phosphate as a signaling molecule and its sensing mechanism. Physiol Rev. 2018;98(4):2317–2348. doi:10.1152/physrev.00022.2017

110.

Pandol

. Serum phosphate levels and alcohol-induced pancreatitis. Gastroenterology. 2022;162(3):995–996. doi:10.1053/j.gastro.2021.11.018

111.

Henaut

Massy

. Phosphate meeting cholesterol-consequences for cardiovascular disease in chronic kidney disease?. Kidney Int. 2021;99(6):1264–1267. doi:10.1016/j.kint.2021.02.022

112.

Dhingra

Gona

Benjamin

, et al. Relations of serum phosphorus levels to echocardiographic left ventricular mass and incidence of heart failure in the community. Eur J Heart Fail. 2010;12(8):812–818. doi:10.1093/eurjhf/hfq106

113.

Tonelli

Sacks

Pfeffer

, et al. Relation between serum phosphate level and cardiovascular event rate in people with coronary disease. Circulation. 2005;112(17):2627–2633. doi:10.1161/CIRCULATIONAHA.105.553198

114.

Bortnick

Kim

, et al. Biomarkers of mineral metabolism and progression of aortic valve and mitral annular calcification: The multi-ethnic study of atherosclerosis. Atherosclerosis. 2019;285:79–86. doi:10.1016/j.atherosclerosis.2019.04.215

115.

Müller

Bertsch

Volke

, et al. Narrative review of metabolomics in cardiovascular disease. J Thorac Dis. 2021;13(4):2532–2550. doi:10.21037/jtd-21-22

116.

Moskaleva

Shestakova

Kukharenko

, et al. Target metabolome profiling-based machine learning as a diagnostic approach for cardiovascular diseases in adults. Metabolites. 2022;12(12):1185. doi:10.3390/metabo12121185

117.

Polidori

Grasso

Chiarelli

, et al. Amino acid-related metabolic signature in obese children and adolescents. Nutrients. 2022;14(7). doi:10.3390/nu14071454

118.

Hahn

Selvaraj

Sharma

, et al. Towards metabolomic-based precision approaches for classifying and treating heart failure. JACC Basic Transl Sci. 2024;9(9):1144–1158. doi:10.1016/j.jacbts.2024.04.008

119.

Culler

Sinha

Filipp

, et al. Metabolomic profiling identifies novel metabolites associated with cardiac dysfunction. Sci Rep. 2024;14(1):20694. doi:10.1038/s41598-024-71329-y

120.

Fernandes Silva

Laakso

. Advances in metabolomics: A comprehensive review of type 2 diabetes and cardiovascular disease interactions. Int J Mol Sci. 2025;26(8). doi:10.3390/ijms26083572

121.

Zieleniewska

Jamiołkowski

Kondraciuk

, et al. Preclinical atherosclerosis and prediabetes: A cross-sectional metabolic assessment in apparently healthy population. Cardiovasc Diabetol. 2025;24(1):280. doi:10.1186/s12933-025-02841-2

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

1.97 MB

1.46 MB

1.23 MB

2.18 MB

4.17 MB

0.13 MB

0.00 MB

3.70 MB