Metabolic Biomarkers for Peripheral Artery Disease Compared with Coronary Artery Disease: Lipoprotein and metabolite profiling of 31,657 individuals from five prospective cohorts

Background Peripheral artery disease (PAD) and coronary artery disease (CAD) represent atherosclerosis in different vascular beds. We conducted detailed metabolic profiling to identify biomarkers for the risk of developing PAD and compared with risk of CAD to explore common and unique risk factors for these different vascular diseases. Methods We measured blood biomarkers using nuclear magnetic resonance metabolomics in five Finnish prospective general-population cohorts (FINRISK 1997, 2002, 2007, 2012, and Health 2000 studies, n = 31 657). We used Cox modelling to estimate associations between biomarkers and incident symptomatic PAD and CAD (498 and 2073 events, respectively) during median follow-up time of 14 years. Results The pattern of biomarker associations for incident PAD deviated from that for CAD. Apolipoproteins and cholesterol measures were robustly associated with incident CAD (for example, age- and sex-adjusted hazard ratio per SD for higher apolipoprotein B/A-1 ratio: 1.30; 95% confidence interval 1.25-1.36), but not with incident PAD (1.04; 0.95-1.14; P heterogeneity <0.001). Low-density lipoprotein (LDL) particle concentrations were also associated with incident CAD (e.g. small LDL particles: 1.24; 1.19-1.29) but not with PAD (1.07; 0.98-1.17; P heterogeneity <0.001). In contrast, more consistent associations of smaller LDL particle size and higher triglyceride levels in LDL and HDL particles with increased risk for both CAD and PAD events were seen (P heterogeneity >0.05). Many non-traditional biomarkers, including fatty acids, amino acids, inflammation- and glycolysis-related metabolites were associated with future PAD events. Lower levels of linoleic acid, an omega-6 fatty acid, and higher concentrations of glucose, lactate, pyruvate, glycerol and glycoprotein acetyls were more strongly associated with incident PAD as compared to CAD (Pheterogeneity<0.001). The differences in metabolic biomarker associations for PAD and CAD remained when adjusting for body mass index, smoking, prevalent diabetes, and medications. Conclusions The metabolic biomarker profile for future PAD risk is largely distinct from that of CAD. This may represent pathophysiological differences and may facilitate risk prediction.