Metabolic Effects of the SGLT2 Inhibitor Dapagliflozin in Heart Failure Across the Spectrum of Ejection Fraction

Abstract

Background: Mechanisms of benefit with SGLT2is (sodium-glucose cotransporter-2 inhibitors) in heart failure (HF) remain incompletely characterized. Dapagliflozin alters ketone and fatty acid metabolism in HF with reduced ejection fraction though similar effects have not been observed in HF with preserved ejection fraction. We explore whether metabolic effects of SGLT2is vary across the left ventricular ejection fraction spectrum and their relationship with cardiometabolic end points in 2 randomized trials of dapagliflozin in HF.

Methods: Metabolomic profiling of 61 metabolites was performed in 527 participants from DEFINE-HF (Dapagliflozin Effects on Biomarkers, Symptoms and Functional Status in Patients With HF With Reduced Ejection Fraction) and PRESERVED-HF (Dapagliflozin in PRESERVED Ejection Fraction HF; 12-week, placebo-controlled trials of dapagliflozin in HF with reduced ejection fraction and HF with preserved ejection fraction, respectively). Linear regression was used to assess changes in principal components analysis-defined metabolite factors with treatment from baseline to 12 weeks, as well as the relationship between changes in metabolite clusters and HF-related end points.

Results: The mean age was 66±11 years, 43% were female, and 33% were self-identified as Black. Two principal components analysis-derived metabolite factors (which were comprised of ketone and short-/medium-chain acylcarnitines) increased with dapagliflozin compared with placebo. Ketosis (defined as 3-hydroxybutyrate >500 μM) was achieved in 4.5% with dapagliflozin versus 1.2% with placebo (P=0.03). There were no appreciable treatment effects on amino acids, including branched-chain amino acids. Increases in several acylcarnitines were consistent across LVEF (P_interaction>0.10), whereas the ketogenic effect diminished at higher LVEF (P_interaction=0.01 for 3-hydroxybutyrate). Increases in metabolites reflecting mitochondrial dysfunction (particularly long-chain acylcarnitines) and aromatic amino acids and decreases in branched-chain amino acids were associated with worse HF-related outcomes in the overall cohort, with consistency across treatment and LVEF.

Conclusions: SGLT2is demonstrate common (fatty acid) and distinct (ketogenic) metabolic signatures across the LVEF spectrum. Changes in key pathways related to fatty acid and amino acid metabolism are associated with HF-related end points and may serve as therapeutic targets across HF subtypes.

Registration: URL: https://www.clinicaltrials.gov; Unique Identifiers: NCT03030235 and NCT02653482.

Keywords: fatty acid; heart failure; ketone bodies; metabolomics; quality of life.

Figure 1:. Box-and-whisker plots of changes in metabolite levels by treatment group.

Changes between baseline and follow-up among individual metabolites heavily loaded on factors 7 and 8 that were differentially associated with treatment arm are detailed using boxplots, stratified by treatment group. The box depicts the 25^th, 50^th, and 75^th percentiles. The upper whisker extends from the 75^th percentile to the largest value no further than 1.5*interquartile range from the hinge. The lower whisker extends from the 25^th percentile to the smallest value at most 1.5*interquartile of the hinge. Data beyond the end of the whiskers (outliers) are not depicted. P-value shown for difference between groups. See Table S1 for metabolite abbreviations.

Figure 2:. Treatment effects of dapagliflozin compared with placebo across the range of ejection fraction for change in select metabolites.

Estimated difference of dapagliflozin compared with placebo and 95% confidence intervals on log-transformed β-hydroxybutyrate (A) and C2 acylcarnitine (B) levels across ejection fraction depicted using restricted cubic spline modeling. P-value for left ventricular ejection fraction*treatment interaction term is shown.

Figure 3:. Heatmap of correlation between changes in metabolites and changes in trial endpoints.

A heatmap cluster correlation matrix (using Pearson correlation) is depicted for the relationship between changes in several trial endpoints (KCCQ-OSS, KCCQ-CSS, log-transformed NT-proBNP, weight, and 6-minute walk test distance) with changes in metabolites from baseline to follow-up. See Supplementary Table 1 for metabolite abbreviations. CSS, clinical summary score; KCCQ, Kansas City Cardiomyopathy Questionnaire; NT-proBNP, N-terminal pro-B-type natriuretic peptide; OSS, overall summary score.

Figure 4:. Metabolomic Profiling of Dapagliflozin Across the Spectrum of Ejection Fraction in Patients with Heart Failure

A total of 527 participants had available metabolite profiling during follow-up. Compared with placebo, dapagliflozin increased metabolite clusters enriched in short/medium-chain acylcarnitines and ketone-related metabolites. The effect of dapagliflozin on biomarkers of fatty acid metabolism was consistent across LVEF, but the ketogenic effect diminished at higher LVEF. In both treatment arms, several metabolites were predictive of heart-failure related endpoints. Created with BioRender.com. AA, amino acids; BCAA, branched-chain amino acids; LCAC, long-chain acylcarnitines; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; LVEF, left ventricular ejection fraction; KCCQ, Kansas City Cardiomyopathy Questionnaire; NT-proBNP, N-terminal pro-B-type natriuretic peptide.