Integrative metabolomics differentiate coronary artery disease, peripheral artery disease, and venous thromboembolism risks

Abstract

Rationale: Arterial and venous cardiovascular conditions, such as coronary artery disease (CAD), peripheral artery disease (PAD), and venous thromboembolism (VTE), are genetically correlated. Interrogating distinct and overlapping mechanisms may shed new light on disease mechanisms.

Objective: In this study, we aimed to: identify and compare (1) epidemiologic and (2) causal, genetic relationships between metabolites and CAD, PAD, and VTE.

Methods: We used metabolomic data from 95,402 individuals in the UK Biobank, excluding individuals with prevalent cardiovascular disease. Logistic regression models adjusted for age, sex, genotyping array, first five principal components of ancestry, and statin use estimated the epidemiologic associations of 249 metabolites with incident CAD, PAD, or VTE. Bidirectional two-sample Mendelian randomization (MR) estimated the causal effects between metabolites and cardiovascular phenotypes using genome-wide association summary statistics for metabolites (N = 118,466 from UK Biobank), CAD (N = 184,305 from CARDIoGRAMplusC4D 2015), PAD (N = 243,060 from Million Veterans Project) and VTE (N = 650,119 from Million Veterans Project). Multivariable MR (MVMR) was performed in subsequent analyses.

Results: We found that 194, 111, and 69 metabolites were epidemiologically associated (P < 0.001) with CAD, PAD, and VTE, respectively. Metabolomic profiles exhibited variable similarity between disease pairs: CAD and PAD (N = 100 shared associations, R² = 0.499), CAD and VTE (N = 68, R² = 0.455), and PAD and VTE (N = 54, R² = 0.752). MR revealed 28 metabolites that increased risk for both CAD and PAD and 2 metabolites that increased risk for CAD but decreased risk for VTE. Despite strong epidemiologic overlap, no metabolites had a shared genetic relationship between PAD and VTE. MVMR revealed several metabolites with shared causal effects on CAD and PAD related to cholesterol content within very-low-density lipoprotein particles.

Conclusions: While common arterial and venous conditions are associated with overlapping metabolomic profiles, MR prioritized the role of remnant cholesterol in arterial diseases but not venous thrombosis.

Figure 1.

Study schematic. (A) Epidemiological association analysis was performed using individual-level data from the UK Biobank, regressing incident coronary artery disease (CAD)/peripheral artery disease (PAD)/venous thromboembolism (VTE) events on metabolite levels. (B) Bidirectional Mendelian randomization analysis was performed using metabolite summary statistics from the UK Biobank as exposures and CAD/PAD/VTE summary statistics from either CARDIoGRAMplusC4D 2015 or Million Veterans Program (MVP) as outcomes, and (C) vice versa. (D) Taking only the metabolite with significant causal effects from (B), multivariable Mendelian randomization analysis was performed using metabolite summary statistics from the UK Biobank as exposures and CAD/PAD/VTE summary statistics from either CARDIoGRAMplusC4D 2015 or MVP as outcomes.

Figure 2.

Epidemiological associations between metabolites and incident cardiovascular outcomes. Associations were estimated using logistic regression on 95,402 individuals in the UK Biobank and 7 classes of metabolites: VLDL = very low-density lipoprotein, FA = fatty acid, HDL = high-density lipoprotein, IDL = intermediate-density lipoprotein, LDL = low-density lipoprotein, M = miscellaneous, NL = non-lipid metabolite, and OL = other lipid metabolite. Bars oriented upwards indicate a positive association while those oriented inwards indicate a negative association. Error bars represent 95% confidence intervals. Transparent bars represent associations that were not significant based on a multiple testing correction of P < 0.05 / 41. Coronary artery disease (CAD) was associated with 194 metabolites. Peripheral artery disease (PAD) was associated with 111 metabolites. Venous thromboembolism (VTE) was associated with 69 metabolites.

Figure 3.

Venn diagram of metabolites with shared significant associations between coronary artery disease (CAD), peripheral artery disease (PAD), and venous thromboembolism (VTE) estimated using logistic regression. Associations were estimated using logistic regression on 95,402 individuals in the UK Biobank. Metabolites that were associated with more than one cardiovascular phenotype but demonstrated an opposite direction of effect were indicated with white asterisks.

Figure 4.

Forest plots of estimated causal effects of genetically-determined metabolites on cardiovascular outcomes estimated using bidirectional Mendelian randomization. Bidirectional, two-sample inverse variance weighted Mendelian randomization using summary statistics from the UK Biobank and CARDIoGRAMplusC4D 2015 or Million Veterans Program (MVP) was performed to estimate causal effects of metabolites on coronary artery disease (CAD), peripheral artery disease (PAD), and venous thromboembolism (VTE). Bars are effect estimates from Mendelian randomization. Error bars represent 95% confidence intervals. Transparent bars represent associations that were not significant based on a multiple testing correction of P < 0.05 / 41. There were 87, 36, and 2 metabolites with a causal effect on CAD, PAD, and VTE, respectively.

Figure 5.

Venn diagram of metabolites with shared causal effects between coronary artery disease (CAD), peripheral artery disease (PAD), and venous thromboembolism (VTE) estimated using bidirectional Mendelian randomization. Bidirectional, two-sample inverse variance weighted Mendelian randomization using summary statistics from the UK Biobank and CARDIoGRAMplusC4D 2015 or Million Veterans Program (MVP) was performed to estimate causal effects of metabolites on CAD, PAD, and VTE. Metabolites that had causal effects on more than one cardiovascular phenotype but demonstrated an opposite direction of effect were indicated with white asterisks.