Abstract
Background:
Heart failure (HF) is heterogeneous syndrome with persistently high mortality. Nuclear Magnetic Resonance (NMR) spectroscopy enables high-throughput metabolomics, suitable for precision phenotyping. We aimed to use targeted metabolomics to derive a metabolic risk score (MRS) that improved mortality risk stratification in HF.
Methods:
NMR was used to measure 21 metabolites (lipoprotein subspecies, branched-chain amino acids, alanine, GlycA, ketone bodies, glucose, and citrate) in plasma collected from a HF community cohort. The MRS was derived using LASSO penalized Cox regression and temporal validation. The association between the MRS and mortality and whether risk stratification was improved over the Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) clinical risk score and N-terminal pro-B-type natriuretic peptide (NT-proBNP) levels were assessed.
Results:
The study included 1382 patients (median age 78 years, 52% men, 43% reduced ejection fraction) with a 5-year survival rate of 48% (95% CI, 46–51%). The MRS included 9 metabolites measured. In the validation dataset, a 1 standard deviation increase in the MRS was associated with a large increased rate of death (HR=2.2; 95% CI 1.9–2.5) that remained after adjustment for MAGGIC score and NT-proBNP (HR=1.6; 95% CI 1.3–1.9). These associations did not differ by ejection fraction. The integrated discrimination and net reclassification indices, and Uno’s C statistic, indicated that the addition of the MRS improved discrimination over MAGGIC and NT-proBNP.
Conclusions:
This MRS developed in a HF community cohort was associated with a large excess risk of death and improved risk stratification beyond an established risk score and clinical markers.
Keywords: biomarkers, heart failure, NMR, metabolomics
Introduction
Heart failure (HF) is a heterogeneous syndrome, which despite recent therapeutic progress is associated with persistently high mortality.1–4 The staggering clinical and public health burden of HF imparts a sense of urgency to elucidate the mechanistic underpinnings of the diverse forms of the HF syndrome.1–4 In clinical practice, HF is categorized according to the ejection fraction (EF), which does not fully account for the complexity of the HF syndrome.5 The 2022 AHA/ACC/HFSA guidelines suggest multi-omics may be key to unraveling the mechanistic underpinnings of the HF syndrome and improving prognosis.1 Specifically, metabolomics, which is the study of small molecules or metabolites in biological samples, has shown promise in HF. However, few studies have examined the use of metabolomics and HF mortality.6–9
Nuclear magnetic resonance (NMR) spectroscopy has been used to study metabolomics in cardiovascular disease10 and shown high comparability with conventional chemistry measurements.11Specifically, the Vantera® NMR Clinical Analyzer is a clinically deployed high-throughput targeted platform, on which a metabolomics assay suitable for large epidemiologic studies was developed.10, 12–16 This targeted metabolomics assay can measure several classes of metabolites from stored blood samples, including a standard lipid panel, lipoprotein particles, ketone bodies, branched-chain amino acids, additional small molecule metabolites, and a marker of systemic inflammation. We chose to evaluate a commercially available assay capable of measuring several metabolites from a single blood sample, because of its clinical applicability.
Herein, we used targeted metabolomics to develop and validate a metabolic risk score (MRS) to identify novel signatures of the HF syndrome and evaluate their associations with mortality. In doing so, our overarching goal was to examine whether the addition of biomarkers from targeted metabolomics might improve risk stratification in patients with HF, beyond guideline recommended risk stratification tools,1 such as Meta-analysis Global Group in Chronic HF (MAGGIC) scores and natriuretic peptides.
Methods
The data that support the findings of this study are available from the corresponding author upon reasonable request. The Mayo Clinic and Olmsted Medical Center Institutional Review Boards approved of this study and informed consent was obtained from each patient. Full methods are now available as supplemental data, including Supplemental Table 1.
Results
After excluding 7 patients with insufficient plasma volume, NMR measurements were available for 1382 out of 1389 patients enrolled in our HF community cohort. The overall 5-year survival rate was 48% (95% CI 46–51%). Patients were elderly (median age of 78 years), white (97%) and approximately equally distributed between men (52%) and women (48%) (Table 1). The median (interquartile range) MAGGIC score was 24 (20–29), NT-proBNP level was 8899 (4184,16351) pg/ml, and 43% of the patients had an EF <50%. Most patients were hypertensive and more than a third were diagnosed with diabetes. Distributions of metabolite concentrations are presented in Table 2. Baseline clinical characteristics and metabolite concentrations were largely similar across both the developmental and validation cohorts. One notable exception was the prevalence of atrial fibrillation which was higher in the validation cohort (57%) compared to the development cohort (33%).
Table 1.
Baseline clinical characteristics and outcomes for heart failure community cohort.
| Characteristic | Overall N = 1382 |
Development N = 880 |
Validation N = 502 |
|---|---|---|---|
| Demographics and Medical History | |||
| Age (years) | 78 (68, 84) | 78 (69, 84) | 76 (66, 84) |
| Male sex | 715 (52) | 438 (50) | 277 (55) |
| White race | 1334 (97) | 851 (97) | 483 (96) |
| Body mass index (kg/m2) | 28 (24, 33) | 28 (24, 33) | 29 (25, 34) |
| Current smoker | 144 (10) | 90 (10) | 54 (11) |
| Hypertension | 1261 (91) | 795 (90) | 466 (93) |
| Diabetes | 493 (36) | 297 (34) | 196 (39) |
| COPD | 395 (29) | 270 (31) | 125 (25) |
| Atrial fibrillation | 576 (42) | 288 (33) | 288 (57) |
| Cerebrovascular disease | 395 (29) | 274 (31) | 121 (24) |
| HF Characteristics | |||
| NYHA (Class 3 or 4) | 950 (69) | 600 (68) | 350 (70) |
| Reduced Ejection Fraction (<50%) | 587 (43) | 354 (41) | 233 (48) |
| Ischemic etiology | 697 (50) | 450 (51) | 247 (49) |
| MAGGIC score | 24 (20,29) | 25 (20,29) | 23 (19,27) |
| Laboratory Measurements | |||
| NT-proBNP (pg/ml) | 8899 (4184, 16351) | 9271 (4091, 17117) | 8483 (4255, 15221) |
| Outcomes | |||
| Deaths during follow-up | 1030 (75) | 672 (76) | 358 (71) |
| Mortality rate per 100 person-years (95% CI) | 14.5 (13.6, 15.4) | 15.7 (14.5, 16.8) | 12.7 (11.4, 14.0) |
COPD = Chronic obstructive pulmonary disease; EF=Ejection Fraction; NYHA = New York Heart Association; MAGGIC=Meta-Analysis Global Group in Chronic Heart Failure.
Figures indicate numbers with percentage in parentheses or median (interquartile range), unless otherwise specified.
Table 2.
Concentrations of metabolites expressed as median and interquartile range. The 9 metabolites retained in the metabolic risk score are italicized.
| Overall N=1382 |
Development N=880 |
Validation N=502 |
|
|---|---|---|---|
| TRLP (nmol/L) | |||
| Very Large (90–240 nm) | 0.11 (0.05–0.28) | 0.10 (0.05–0.24) | 0.15 (0.06–0.32) |
| Large (50–89 nm) | 1.37 (0.19–4.19) | 1.10 (0.16–3.75) | 1.87 (0.34–4.92) |
| Medium (37–49 nm) | 12 (5–23) | 10 (4–21) | 15 (6–26) |
| Small (30–36 nm) | 29 (13–49) | 28 (12–47) | 31 (16–52) |
| Very Small (24–29 nm) | 73 (41–108) | 73 (40–111) | 73 (42–106) |
| LDL Particle (nmol/L) | |||
| Large (21.5–23 nm) | 195 (84–339) | 209 (90–354) | 174 (73–315) |
| Medium (20.5–21.4 nm) | 175 (24–361) | 154 (13–339) | 210 (46–393) |
| Small (19–20.4 nm) | 622 (428–892) | 629 (432–894) | 611 (423–883) |
| Calibrated HDL Particle (μmol/L) | |||
| H7P: HDL subspecies (12 nm) | 0.25 (0.08–0.55) | 0.25 (0.07–0.56) | 0.25 (0.10–0.52) |
| H6P: HDL subspecies (10.8 nm) | 0.50 (0.10–1.20) | 0.50 (0.08–1.20) | 0.48 (0.13–1.25) |
| H5P: HDL subspecies (10.3 nm) | 0.45 (0.12–0.94) | 0.42 (0.10–0.87) | 0.56 (0.15–1.06) |
| H4P: HDL subspecies (9.5 nm) | 1.83 (1.17–2.67) | 1.93 (1.21–2.76) | 1.68 (1.14–2.53) |
| H3P: HDL subspecies (8.7 nm) | 1.33 (0.35–2.80) | 1.37 (0.41–2.91) | 1.25 (0.30–2.59) |
| H2P: HDL subspecies (7.8 nm) | 7.7 (5.0–10.1) | 7.1 (4.5–9.6) | 8.7 (6.4–10.8) |
| H1P: HDL subspecies (7.4 nm) | 0.15 (0.00–1.66) | 0.00 (0.00–1.28) | 0.44 (0.00–2.18) |
| Amino Acids (μmol/L) | |||
| BCAA (Valine, Leucine, Isoleucine) | 408 (340–480) | 410 (342–476) | 403 (338–487) |
| Alanine | 410 (334–492) | 390 (317–472) | 447 (371–526) |
| GlycA (μmol/L) | 453 (391–537) | 463 (398–553) | 439 (386–510) |
| Ketone Bodies (μmol/L) | 180 (134–308) | 191 (137–354) | 167 (128–250) |
| Small Molecule Metabolites | |||
| Glucose (mg/dL) | 105 (91–128) | 106 (91–132) | 103 (90–125) |
| Citrate (μmol/L) | 121 (97–149) | 118 (94–146) | 125 (104–155) |
TRLP = Triglyceride Rich Lipoprotein Particle; LDL = Low-Density Lipoprotein; HDL = High-Density Lipoprotein; BCAA = Branched-Chain Amino Acids
Metabolomic Risk Score Development and Validation
In the development data set, 11 metabolites were associated with mortality, of which 9 were selected by the LASSO model to generate MRS. The corresponding beta coefficients from our MRS are summarized in Figure 1. For further context age- and sex- adjusted HRs for the 9 markers are summarized in Supplemental Table 2. Six markers (H1P, H2P, M- and S-TRLP, BCAA and alanine) were negatively associated with mortality risk, whereas three markers (H6P, glucose and ketone bodies) were positively associated with mortality risk. Survival differed markedly across the three MRS risk groups (Figure 2A). The cumulative survival at 5 years in the low, intermediate, and high MRS risk groups was 71% (95% CI 65–77%), 43% (95% CI 39–48%) and 20% (95% CI 15–26%) respectively.
Figure 1.
β coefficients of the 9 markers in the metabolic risk score.
BCCA = Branched-Chain Amino Acids; TRLP = Triglyceride Rich Lipoprotein Particle; H1P = HDLP subspecies (7.4 nm); H2P = HDLP subspecies (7.8 nm); H6P = HDLP subspecies (10.8 nm).
Figure 2.
Kaplan-Meier survival curves by metabolic risk score groups in the development (A) and validation (B) data sets (low risk in blue, intermediate risk in yellow and high risk in black). Multivariable adjusted hazard ratios (HRs) and 95% confidence intervals per 1 standard deviation increase in metabolic risk score in the development (C) and validation (D) data sets.
MAGGIC = Meta-Analysis Global Group in Chronic Heart Failure clinical risk score; MRS = metabolic risk score; NT-proBNP = N-terminal pro-B-type natriuretic peptide.
In the validation data set, a 1 standard deviation increase in MRS was associated with a more than twofold excess rate of death (HR=2.2; 95% CI 1.9–2.5) and was slightly attenuated after adjustment for the MAGGIC score and NT-proBNP (HR=1.6; 95% CI 1.3–1.9) (Figure 2D). Similar results were observed when square root transformation was applied to the measurements. These associations were consistent across EF categories with no significant interactions observed (Supplemental Figure 1). The MRS distribution was similar among ischemic and non-ischemic HF patients (P-value = .7), with median (IQR) values of 0.04 (−0.42, 0.45) and 0.02 (−0.49, 0.50), respectively. We noted similar improvements in performance measures of the MRS across among ischemic and non-ischemic HF patients. The excess rate of death remained when restricted to cardiovascular disease-related mortality in crude and MAGGIC-adjusted models (subdistribution HR=2.0; 95% CI 1.6–2.5 and HR=1.4; 95% CI 1.1–1.9, respectively) but was attenuated after further adjustment for NT-proBNP (subdistribution HR=1.2; 95% CI 0.9–1.6).
Incremental Value of MRS Beyond Established Clinical Factors
We examined the continuous net reclassification improvement and integrated discrimination improvement at 5 years and found the addition of MRS improved model performance beyond these clinical variables (Table 3). Similar results were observed at 1 and 3 years (Supplemental Table 3). The Uno’s C statistic in the validation data set indicated that the addition of the MRS improved discrimination over MAGGIC and NT-proBNP (Figure 3); that improvement was only statistically significant at 1 and 3 years (Supplemental Figure 2).
Table 3.
Incremental benefit of metabolic risk score at 5 years. The numbers indicate the percent change in integrated discrimination improvement and the continuous net reclassification improvement in the validation dataset.
| Integrated Discrimination Improvement | Continuous Net Reclassification Improvement | |||
|---|---|---|---|---|
| Base Model | Improvement | 95% CI | Improvement | 95% CI |
| MAGGIC | 3.5% | 0.8–6.0% | 14.4% | 4.6–23.0% |
| MAGGIC + NT-proBNP | 2.4% | 0.4–4.4% | 13.4% | 2.1–20.5% |
MAGGIC = Meta-Analysis Global Group in Chronic Heart Failure clinical risk score; NT-proBNP = N-terminal pro-B-type natriuretic peptide.
Figure 3.
Metabolic risk score C-statistic compared to that of MAGGIC and NT-proBNP alone or in combination in the validation data set.
MAGGIC = Meta-Analysis Global Group in Chronic Heart Failure clinical risk score; MRS = metabolic risk score; NT-proBNP = N-terminal pro-B-type natriuretic peptide.
Discussion
Using targeted metabolomics, we developed and validated a novel MRS for mortality risk in HF syndrome. The MRS was associated with a large excess risk of mortality in HF and this strong association was independent of the MAGGIC score, recommended by the 2022 HF guideline for clinical risk stratification tools.1 The association with mortality persisted across EF categories and after further adjustment for NT-proBNP. By relying on a community cohort to conduct these analyses, our data adds to the weight of evidence that circulating metabolites improves risk stratification across the entire spectrum of the HF syndrome.
Previous studies of metabolomics in HF
Metabolomics can be used to assess changes in a variety of metabolite classes potentially important in HF, including amino acids, fatty acids, and various organic compounds. However, little is known about how metabolomics can be leveraged to identify unique metabolic signatures in patients with HF to assess their association with clinical outcomes including mortality.
Comparing our results to previously published studies is challenging due to the heterogeneity of the populations studied (in some cases restricted to HF with reduced or with preserved EF), measurement platforms (NMR or mass spectrometry) and outcome definitions.6–9 Further, several published studies contained relatively small sample sizes.7–9
The current approach relied on targeted metabolomics, whereby a preselected group of metabolites is measured quantitatively,17 to derive a MRS associated with mortality. NMR-based risk scores have been developed using a similar NMR LipoProfile analysis approach, from different populations of individuals at risk of cardiovascular disease, not specifically HF.18
Most similar to our approach, Lanfear et al. used a targeted approach to measure 87 metabolites reduced to 13 by penalized regression from which they derived a risk score validated by randomly splitting the cohort.6 Their study was restricted to HF patients with reduced EF and reported a strong association between their MRS and mortality which was only partially attenuated by adjustment for known clinical parameters.
Our findings augment this body of knowledge as we analyzed a large cohort that included all forms of HF, regardless of EF, and represented the comprehensive experience of the community of patients living with HF. Using a different platform but similar analytic approach, we derived and validated our MRS using 9 metabolites. Notably, our findings did not differ according to the type of HF (e.g., reduced or preserved EF) and branch chain amino acids were also retained in our MRS. The similarities between these data bring an important dimension of congruency.
Specific metabolites and biological plausibility
Although the primary goal of our study was to examine the utility of targeted metabolomics in HF risk stratification rather than identify novel biomarkers, it is important to examine biological plausibility of the markers identified in our MRS. Positive and negative beta coefficients were observed for the 9 metabolites retained in our MRS. These results suggest that some metabolites may be associated with an increased risk of death, while some may be protective. Notably, we observed smaller high-density lipoprotein particles (H2P, H1P) provided a protective contribution to our MRS. These findings align with clinical trials of HF and clinically referred populations of HF, spanning the entire EF continuum, showing potential protective effects of small high-density lipoproteins in HF, including increased survival.19, 20
We also found branched-chain amino acids (BCAAs; valine, leucine, and isoleucine) provided a protective contribution to our MRS. BCAAs are essential amino acids associated with protein metabolism in the heart and skeletal muscles, which have been studied in muscle wasting disorders.21, 22 Previous studies using different panels of metabolites similarly identified BCAAs as key risk indicators,6 leading to hypothesize that these associations could reflect the systemic skeletal muscle response in advanced HF states. Our group previously reported on an association between the Metabolic Malnutrition Index, which includes BCAAs, and an increased risk of mortality in this HF community cohort.23 The direct relationship with circulating BCAAs and mortality in patients with HF is less clear. Some studies suggest higher BCAAs increase the risk of mortality24, 25, while others have shown higher BCAAs may decrease risk of mortality.26 Future studies are needed to further elucidate the role of BCAAs in HF mortality.
Conversely, glucose was associated with an increased MRS in our study. Elevated glucose levels are a hallmark of diabetes. Diabetes is a well-known risk factor for cardiovascular mortality, including HF.27, 28 Studies have found hyperglycemia is a predictor of mortality in patients with HF.29 Collectively, these results support the biological plausibility of the metabolites retained in our MRS. Further, these results highlight a major strength of targeted metabolomics in HF risk stratification by capturing several potential domains of HF mortality risk rather than relying on a single biomarker.
Clinical implications
Published scores for risk assessment often underperform when deployed in real world practice.30, 31 While guideline recommended tools to stratify risk in HF (i.e., MAGGIC and natriuretic peptides) performed similarly in our community cohort compared to what has been reported in clinically selected populations,32–34 the improved discriminatory ability of the MRS above these tools emphasizes its clinical relevance. Indeed, novel risk stratification tools, such as metabolomics, have the potential to be clinical helpful in non-selected populations encompassing the entire HF spectrum.
Limitations and Strengths
Our study has several limitations to consider when evaluating the results. Our cohort was predominantly of European ancestry, limiting our ability to assess generalizability of our findings to other populations. A targeted panel was used to develop the MRS which may not have included all the metabolites that are predictive of mortality in patients with HF.
Our study has several important strengths. Firstly, using an NMR analyzer already employed for routine clinical testing, we studied a community-based cohort, and thus our results are widely applicable and relevant to patients living with HF in the community with minimal referral bias. Secondly, nearly all in-patient and outpatient encounters are captured by the linkage system of the Rochester Epidemiology Project, which results in a uniquely comprehensive set of clinical information with complete follow-up. Lastly, the use of rigorous statistical methods strengthens the inference that can be drawn from the results. To this end, we relied on a rigorous analytical approach including temporal validation, as recommended by the transparent reporting of a multivariable prediction model for individual prognosis or diagnosis guideline.35
Conclusion
The novel MRS that we identified exhibited a strong association with mortality, markedly augmenting the prognostic information conveyed by existing clinical tools recommended by the guideline such as the MAGGIC score and NT-proBNP. Importantly, the results were independent of the EF. Hence, the novel metabolic markers offer the promise of conveying important risk stratification information across the whole spectrum of the HF syndrome.
Supplementary Material
Acknowledgments:
The authors thank Dr. Mary Walter and Yuhai Dai at NIDDK Clinical Laboratory Core, NIH, USA for their assistance measuring key laboratory variables.
Sources of Funding:
The investigators were supported by the Intramural Research Program of the National Heart, Lung, and Blood Institute of the National Institutes of Health (ZIAHL006278). This study used the resources of the Rochester Epidemiology Project medical records linkage system, which is supported by the National Institute on Aging (NIA; AG 058738), by the Mayo Clinic Research Committee, and by fees paid annually by Rochester Epidemiology Project users. The content of this article is solely the responsibility of the authors and does not represent the official views of the National Institutes of Health or the Mayo Clinic.
Nonstandard Abbreviations and Acronyms
- BCAA
Branched-chain amino acids
- EF
Ejection fraction
- HF
Heart failure
- MAGGIC
Meta-Analysis Global Group in Chronic Heart Failure
- MRS
Metabolic risk score
- NMR
Nuclear Magnetic Resonance
- NT-proBNP
N-Terminal pro-B Type Natriuretic Peptide
Footnotes
Disclosures: MAC is an employee of and holds stock in Labcorp. JDO is a consultant, stockholder, and former employee of Labcorp.
References:
- 1.Heidenreich PA, Bozkurt B, Aguilar D, Allen LA, Byun JJ, Colvin MM, Deswal A, Drazner MH, Dunlay SM, Evers LR, et al. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2022;145:e895–e1032. [DOI] [PubMed] [Google Scholar]
- 2.Ziaeian B and Fonarow GC. Epidemiology and aetiology of heart failure. Nat Rev Cardiol. 2016;13:368–78. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Roger VL. Epidemiology of Heart Failure: A Contemporary Perspective. Circ Res. 2021;128:1421–1434. [DOI] [PubMed] [Google Scholar]
- 4.Toward Precision Medicine: Building a Knowledge Network for Biomedical Research and a New Taxonomy of Disease Washington (DC); 2011. [PubMed] [Google Scholar]
- 5.Francis GS, Cogswell R and Thenappan T. The heterogeneity of heart failure: will enhanced phenotyping be necessary for future clinical trial success? J Am Coll Cardiol. 2014;64:1775–6. [DOI] [PubMed] [Google Scholar]
- 6.Lanfear DE, Gibbs JJ, Li J, She R, Petucci C, Culver JA, Tang WHW, Pinto YM, Williams LK, Sabbah HN and Gardell SJ. Targeted Metabolomic Profiling of Plasma and Survival in Heart Failure Patients. JACC Heart Fail. 2017;5:823–832. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Kouzu H, Katano S, Yano T, Ohori K, Nagaoka R, Inoue T, Takamura Y, Ishigo T, Watanabe A, Koyama M, et al. Plasma amino acid profiling improves predictive accuracy of adverse events in patients with heart failure. ESC Heart Fail. 2021;8:5045–5056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Zhao H, Shui B, Zhao Q, Hu Z, Shu Q, Su M, Zhang Y and Ni Y. Quantitative Metabolomics Reveals Heart Failure With Midrange Ejection Fraction as a Distinct Phenotype of Heart Failure. Can J Cardiol. 2021;37:300–309. [DOI] [PubMed] [Google Scholar]
- 9.Vignoli A, Fornaro A, Tenori L, Castelli G, Cecconi E, Olivotto I, Marchionni N, Alterini B and Luchinat C. Metabolomics Fingerprint Predicts Risk of Death in Dilated Cardiomyopathy and Heart Failure. Front Cardiovasc Med. 2022;9:851905. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Rankin NJ, Preiss D, Welsh P, Burgess KE, Nelson SM, Lawlor DA and Sattar N. The emergence of proton nuclear magnetic resonance metabolomics in the cardiovascular arena as viewed from a clinical perspective. Atherosclerosis. 2014;237:287–300. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Petersen M, Dyrby M, Toubro S, Engelsen SB, Norgaard L, Pedersen HT and Dyerberg J. Quantification of lipoprotein subclasses by proton nuclear magnetic resonance-based partial least-squares regression models. Clin Chem. 2005;51:1457–61. [DOI] [PubMed] [Google Scholar]
- 12.Jeyarajah EJ, Cromwell WC and Otvos JD. Lipoprotein particle analysis by nuclear magnetic resonance spectroscopy. Clin Lab Med. 2006;26:847–70. [DOI] [PubMed] [Google Scholar]
- 13.Matyus SP, Braun PJ, Wolak-Dinsmore J, Jeyarajah EJ, Shalaurova I, Xu Y, Warner SM, Clement TS, Connelly MA and Fischer TJ. NMR measurement of LDL particle number using the Vantera Clinical Analyzer. Clin Biochem. 2014;47:203–10. [DOI] [PubMed] [Google Scholar]
- 14.Andersson C, Liu C, Cheng S, Wang TJ, Gerszten RE, Larson MG and Vasan RS. Metabolomic signatures of cardiac remodelling and heart failure risk in the community. ESC Heart Fail. 2020;7:3707–3715. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Sokooti S, Flores-Guerrero JL, Heerspink HJL, Connelly MA, Bakker SJL and Dullaart RPF. Triglyceride-rich lipoprotein and LDL particle subfractions and their association with incident type 2 diabetes: the PREVEND study. Cardiovasc Diabetol. 2021;20:156. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Sokooti S, Flores-Guerrero JL, Kieneker LM, Heerspink HJL, Connelly MA, Bakker SJL and Dullaart RPF. HDL Particle Subspecies and Their Association With Incident Type 2 Diabetes: The PREVEND Study. J Clin Endocrinol Metab. 2021;106:1761–1772. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Iliou A, Mikros E, Karaman I, Elliott F, Griffin JL, Tzoulaki I and Elliott P. Metabolic phenotyping and cardiovascular disease: an overview of evidence from epidemiological settings. Heart. 2021;107:1123–1129. [DOI] [PubMed] [Google Scholar]
- 18.Otvos JD, Shalaurova I, May HT, Muhlestein JB, Wilkins JT, McGarrah RW 3rd, and Kraus WE. Multimarkers of metabolic malnutrition and inflammation and their association with mortality risk in cardiac catheterisation patients: a prospective, longitudinal, observational, cohort study. Lancet Healthy Longev. 2023;4:e72–e82. [DOI] [PubMed] [Google Scholar]
- 19.Hunter WG, McGarrah RW 3rd, Kelly JP, Khouri MG, Craig DM, Haynes C, Felker GM, Hernandez AF, Velazquez EJ, Kraus WE, et al. High-Density Lipoprotein Particle Subfractions in Heart Failure With Preserved or Reduced Ejection Fraction. J Am Coll Cardiol. 2019;73:177–186. [DOI] [PubMed] [Google Scholar]
- 20.Teis A, Cediel G, Amigo N, Julve J, Aranyo J, Andres-Cordon J, Puig-Jove C, Castelblanco E, Gual-Capllonch F, Ferrer-Sistach E, et al. Particle size and cholesterol content of circulating HDL correlate with cardiovascular death in chronic heart failure. Sci Rep. 2021;11:3141. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Holecek M Branched-chain amino acids in health and disease: metabolism, alterations in blood plasma, and as supplements. Nutr Metab (Lond). 2018;15:33. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hiraiwa H, Okumura T and Murohara T. Amino acid profiling to predict prognosis in patients with heart failure: an expert review. ESC Heart Fail. 2023;10:32–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Conners K, Shearer JJ, Joo J, Park H, Manemann SM, Remaley AT, Otvos JD, Connelly MA, Sampson M, Bielinski SJ, et al. The Metabolic Vulnerability Index: A Novel Marker for Mortality Prediction in Heart Failure. JACC Heart Failure 2023;(accepted for publication). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Du X, Li Y, Wang Y, You H, Hui P, Zheng Y and Du J. Increased branched-chain amino acid levels are associated with long-term adverse cardiovascular events in patients with STEMI and acute heart failure. Life Sci. 2018;209:167–172. [DOI] [PubMed] [Google Scholar]
- 25.Du X, You H, Li Y, Wang Y, Hui P, Qiao B, Lu J, Zhang W, Zhou S, Zheng Y, et al. Relationships between circulating branched chain amino acid concentrations and risk of adverse cardiovascular events in patients with STEMI treated with PCI. Sci Rep. 2018;8:15809. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Moissl AP, Lorkowski S, Meinitzer A, Pilz S, Scharnagl H, Delgado GE, Kleber ME, Kramer BK, Pieske B, Grubler MR, et al. Association of branched-chain amino acids with mortality-the Ludwigshafen Risk and Cardiovascular Health (LURIC) study. iScience. 2023;26:106459. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Dauriz M, Targher G, Laroche C, Temporelli PL, Ferrari R, Anker S, Coats A, Filippatos G, Crespo-Leiro M, Mebazaa A, et al. Association Between Diabetes and 1-Year Adverse Clinical Outcomes in a Multinational Cohort of Ambulatory Patients With Chronic Heart Failure: Results From the ESC-HFA Heart Failure Long-Term Registry. Diabetes Care. 2017;40:671–678. [DOI] [PubMed] [Google Scholar]
- 28.Raghavan S, Vassy JL, Ho YL, Song RJ, Gagnon DR, Cho K, Wilson PWF and Phillips LS. Diabetes Mellitus-Related All-Cause and Cardiovascular Mortality in a National Cohort of Adults. J Am Heart Assoc. 2019;8:e011295. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Cho JY, Kim KH, Lee SE, Cho HJ, Lee HY, Choi JO, Jeon ES, Kim MS, Kim JJ, Hwang KK, et al. Admission Hyperglycemia as a Predictor of Mortality in Acute Heart Failure: Comparison between the Diabetics and Non-Diabetics. J Clin Med. 2020;9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Gerber Y, Weston SA, Enriquez-Sarano M, Jaffe AS, Manemann SM, Jiang R and Roger VL. Contemporary Risk Stratification After Myocardial Infarction in the Community: Performance of Scores and Incremental Value of Soluble Suppression of Tumorigenicity-2. J Am Heart Assoc. 2017;6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Singh M, Reeder GS, Jacobsen SJ, Weston S, Killian J and Roger VL. Scores for post-myocardial infarction risk stratification in the community. Circulation. 2002;106:2309–14. [DOI] [PubMed] [Google Scholar]
- 32.Zhang L, Cunningham JW, Claggett BL, Jacob J, Mendelson MM, Serrano-Fernandez P, Kaiser S, Yates DP, Healey M, Chen CW, et al. Aptamer Proteomics for Biomarker Discovery in Heart Failure With Reduced Ejection Fraction. Circulation. 2022;146:1411–1414. [DOI] [PubMed] [Google Scholar]
- 33.Gui H, She R, Luzum J, Li J, Bryson TD, Pinto Y, Sabbah HN, Williams LK and Lanfear DE. Plasma Proteomic Profile Predicts Survival in Heart Failure With Reduced Ejection Fraction. Circ Genom Precis Med. 2021;14:e003140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Rich JD, Burns J, Freed BH, Maurer MS, Burkhoff D and Shah SJ. Meta-Analysis Global Group in Chronic (MAGGIC) Heart Failure Risk Score: Validation of a Simple Tool for the Prediction of Morbidity and Mortality in Heart Failure With Preserved Ejection Fraction. J Am Heart Assoc. 2018;7:e009594. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Collins GS, Reitsma JB, Altman DG and Moons KG. Transparent Reporting of a multivariable prediction model for Individual Prognosis or Diagnosis (TRIPOD): the TRIPOD statement. Ann Intern Med. 2015;162:55–63. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.