Machine Learning Analysis of Identifies Polyunsaturated Fatty Acid Metabolites Predictive of Adverse Outcomes In Heart Failure with Preserved Ejection Fraction Patients

Abstract

which molecules were associated with future risk of either all cause death or a combined adverse outcome of death or rehospitalization in the setting of HFpEF. Results: In patients with HFpEF, increased levels of 9(10)-Epome, 15(R)-PGE1, 17-oxoRvD1, TXB3, RvD3, 5(S),15(S)-DiHETE, and 11dh-2,3-dinor TXB2 at baseline were predictive of all cause mortality (all p<0.05). Increased baseline levels of 8-oxoRvD1, MaR1(n-3DPA), PGE3, and 5,6-DiHETrE were predictive of the combined adverse outcome of death or rehospitalization (all p<0.05). Conclusion: These findings support the hypothesis that distinct PUFA metabolites play a significant role in mediating cardiovascular disease in HFpEF. Our study introduces a novel lipidomics framework for the diagnostic and prognostic assessment of cardiovascular risk in HFpEF patients.