Abstract Background Large-scale biobanks and advancements in high-throughput technologies have provided extensive epidemiological, genomic, and metabolomic tools to identify molecular risk factors for incident cardiovascular diseases (CVDs). Metabolic signature of prevalent CVDs may additionally provide clinical diagnostic tools and inform on key pathological molecular processes. Purpose We aim to define metabolic signature of 74 prevalent CVDs by assessing their association and potentially causal effects on 25 plasma metabolites including amino acids, ketone bodies, fatty acids and glycolysis markers. Methods Cross-sectional analyses used baseline data from the UK Biobank population (n=231 053). General linear models (GLM) assessed associations between 25 metabolites and 74 ICD10-defined CVDs, adjusting for potential confounders. Sensitivity analyses excluded self-reported and incident cases. Additionally, 6 polygenic risk scores (PRS) for various CVDs were tested for metabolite associations using GLM, adjusting for age, sex, genotyping batch, and genetic ancestry. Two-sample Mendelian randomization (MR) analyses assessed potential causality using independent genetic instruments for various CVDs, obtained from large-scale genome-wide association studies (GWAS), while metabolic GWAS summary statistics were derived from the UK Biobank. Causal estimates were obtained using the inverse-variance weighted (IVW) method, with sensitivity analyses using weighted median and MR-Egger methods. A Bonferroni-corrected threshold of significance (p0.05/(74x25)) was applied. Results Cross-sectional GLM analyses identified 8 CVDs that were significantly associated with the largest number of altered plasma metabolites and had a prevalence rate above 1% (Figure). Most CVDs were associated with a lower level of various fatty acids, with the strongest reduction observed for linoleic acid following diagnosis of myocardial infarction or coronary heart disease (CHD, Figure). PRS analyses confirmed these findings, with the most significant (p2x10-6) association linking essential hypertension (HTN) and CHD to a lower level of omega-6 and linoleic acids, which highly correlated with each other (R=0.95). Importantly, two-sample MR analyses provided evidence for a potentially causal relationship, demonstrating that HTN (B=-0.07 SD of metabolite/log(OR), SE=0.01, p=6.7x10-7) and CAD (B=-0.08 SD of metabolite/log(OR), SE=0.01, p=6.5x10-10) contributed to a lower level of linoleic acid using IVW method. Associations remained significant using sensitivity MR methods. Similar results were observed for omega-6 fatty acids level and when continuous blood pressure level was set as an exposure. Conclusions These findings underscore that circulating omega-6 fatty acids, and particularly linoleic acid, are actively modulated by HTN and CAD. This study provides insights that may inform future diagnostic strategies and guide research into the pathomechanisms underlying CVDs progression.Associations between CVDs & metabolites.
Dziedzic et al. (Sat,) studied this question.