Background: Carotid atherosclerotic plaque and its progression significantly predict cardiovascular events. Traditional clinical factors have limited value in early risk prediction. Metabolomics is an emerging technology that can accurately quantify hundreds to thousands of small metabolites in biofluids, thus providing a powerful tool for early biomarker discovery. Previous studies were largely cross-sectional with a focus on the European population, results of which may not be generalized to other ethnic groups. In addition, causality cannot be resolved in cross-sectional analysis. Objective: To prospectively identify novel metabolic markers predictive of carotid plaque onset and progression in American Indians in the Strong Heart Study (SHS). Methods: This study includes 396 apparently healthy American Indians attending both clinical exams in 2001-2003 and 2006-2009 (average 5.5 years of follow-up). All participants had normal fasting glucose and were free of overt CVD at baseline. Plaque progression was defined as having a higher plaque score at the end of follow-up compared to baseline. Fasting plasma metabolites were detected using untargeted LC-MS. A total of 1,364 matching metabolites that passed stringent QC were included in the current analysis. The prospective association of plaque progression with each metabolite was examined using Cox proportional hazards model, adjusting for sex, site, age, BMI, eGFR, lipids, blood pressure, smoking and alcohol drinking at baseline. The combined metabolic effect was examined using a weighted multi-marker metabolites score comprising of all significant metabolites. Incremental value of novel metabolites over traditional risk factors for risk prediction was tested using the net reclassification improvement index (NRI). Multiple testing was corrected using the q-value method (q< 0.05 was considered statistical significance). Results: Of all 396 participants, 100 developed new plaque or progressed to a high order of atherosclerosis during follow-up. Five metabolites including PC (6:2/14:2) (HR 1.56[1.23-1.96]), proscillaridin A (HR 1.68 [1.27-2.21]), 8e-heptadecenedioic acid (HR 1.56 [1.22-1.98]), flavone (HR 1.69 [1.27-2.26]), and n-palmitoyl methionine (HR 1.72 [1.28-2.32]), significantly increase, whereas one metabolite, PC (18:0/18:1) (HR 0.62 [0.52-0.73]), significantly decreases the risk of plaque progression. A multi-marker score comprising of all six metabolites significantly improves risk prediction beyond traditional risk factors (NRI=0.22, P=1.23x10-5). Conclusions: Altered levels of fasting plasma metabolites significantly predict the risk of plaque onset and progression over and above conventional risk factors. The newly detected metabolites may unravel novel metabolic pathways underlying CVD pathogenesis and could be used as new markers for risk prediction and stratification.
Read full abstract