Abstract 23: Metabolomic Features Associated With Incident Hypertension - Results From Insulin Resistance Atherosclerosis Study

Background: Clinically-defined metabolic traits and specific dietary patterns are associated with risk for hypertension (HTN); however, the precise metabolomic features that underlie these associations are not yet well understood. A more detailed understanding of these metabolomic features may shed new light on the pathogenesis of HTN and identify novel therapeutic targets. Hypothesis: We hypothesized that there is a metabolic signature associated with high risk for incident HTN. Methods: Ninety three metabolites were quantified by liquid chromatography and mass spectrometry in baseline serum from a subset of the Insulin Resistance Atherosclerosis Study (IRAS) participants (N=500). Unsupervised hierarchical cluster analysis (cluster depth = 2, 3, 4) and principal component analysis (PCA) were performed to define distinct metabolomic phenotypes and orthogonal projections of the metabolomics data. Multivariable negative binomial regression models were used to model the association between metabolomic cluster assignments or metabolomics-derived principal components (PCs) with 5-year incident HTN (blood pressure >=140/90 mmHg or new use of anti-HTN medication), after adjusting for age, gender, race, body mass index, baseline systolic blood pressure and insulin resistance measured by intravenous glucose tolerance test. Results: Of 500 individuals, 133 (26%) developed incident HTN. Cluster analysis identified a subset of participants (N= 154, 31%) with a distinct metabolic phenotype associated with increased risk of incident HTN (adjusted RR 1.65; 95% CI 1.21 - 2.25, p = 0.002). This metabolically defined phenotype was characterized by elevated levels of fatty acids involved in de-novo fatty acid synthesis (saturated fatty acids; SFAs, omega 6 polyunsaturated fatty acid; PUFAs) as well as acylcarnitines (involved in fatty acid catabolism). Separately, PC 1 (SFAs and Omega 6 PUFAs) and PC 3 (acylcarnitines) were also significantly associated with incident HTN (Adjusted RR: 1.47, 95% CI 1.25 - 1.73, p <0.0001; and RR 1.20, 95% CI 1.03 - 1.39, p = 0.01). Among SFAs and omega 6 PUFAs, levels of palmitic acid and di-homo-gamma linoleic acid (immediate precursor of arachidonic acid) were the strongest predictors of hypertension (p<0.0001). Acylcarnitine palmitate was the metabolite with strongest association with hypertension in acylcarnitine group (p<0.0001). Conclusion: Hierarchical cluster analysis and PCA uncovered a specific metabolomic signature associated with risk of HTN characterized by elevated levels of metabolites involved in de novo fatty acid synthesis (SFAs and omega 6 PUFAs) and fatty acid catabolism (acylcarnitines). These metabolic perturbation preceded HTN by up to five years. More research is needed to clarify the precise molecular mechanisms that account for this novel association of metabolites with HTN which is independent of the known major risk factors.