If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
First Clinic of Internal Medicine, Department of Internal Medicine, University of Genoa, 16132, Genoa, ItalyIRCCS Ospedale Policlinico San Martino, Largo Rosanna Benzi, 16132, Genoa, Italy
Center for Molecular Cardiology, University of Zurich, 8952, Schlieren, SwitzerlandRoyal Brompton and Harefield Hospitals and Imperial College, London, United KingdomSchool of Cardiovascular Medicine and Sciences, Kings College London, London, UK
In contemporary patients with acute coronary syndromes (ACS), changes in low-density lipoprotein (LDL) electronegativity independently associated with the risk of all-cause and cardiovascular death at 30 days and 1 year, respectively.
•
LDL electronegativity superseded several clinical risk factors for the prediction of 1-year death, including LDL-cholesterol.
•
When added to the GRACE risk score, LDL electronegativity conferred improved discrimination for the prediction of adverse outcomes, emphasizing its predictive utility in this patient population.
•
LDL particles of different electronegativity had a distinct lipidome, with potential implications for their athero-/thrombogenicity.
•
Altered LDL electronegativity is a novel and independent determinant of mortality risk in patients with established atherosclerotic cardiovascular disease over and above the updated GRACE risk score.
Abstract
Background and aims
Low-density lipoprotein (LDL)-cholesterol (LDL-C) promotes atherosclerotic cardiovascular disease (ASCVD), with changes in LDL electronegativity modulating its pro-atherogenic/pro-thrombotic effects. Whether such alterations associate with adverse outcomes in patients with acute coronary syndromes (ACS), a patient population at particularly high cardiovascular risk, remains unknown.
Methods
This is a case-cohort study using data from a subset of 2619 ACS patients prospectively recruited at 4 university hospitals in Switzerland. Isolated LDL was chromatographically separated into LDL particles with increasing electronegativity (L1-L5), with the L1-L5 ratio serving as a proxy of overall electronegativity. Untargeted lipidomics revealed lipid species enriched in L1 (least) vs. L5 (most electronegative subfraction). Patients were followed at 30 days and 1 year. The mortality endpoint was reviewed by an independent clinical endpoint adjudication committee. Multivariable-adjusted hazard ratios (aHR) were calculated using weighted Cox regression models.
Results
Changes in LDL electronegativity were associated with all-cause mortality at 30 days (aHR, 2.13, 95% CI, 1.07–4.23 per 1 SD increment in L1/L5; p=.03) and 1 year (1.84, 1.03–3.29; p=.04), with a notable association with cardiovascular mortality (2.29; 1.21–4.35; p=.01; and 1.88; 1.08–3.28; p=.03). LDL electronegativity superseded several risk factors for the prediction of 1-year death, including LDL-C, and conferred improved discrimination when added to the updated GRACE score (area under the receiver operating characteristic curve 0.74 vs. 0.79, p=.03). Top 10 lipid species enriched in L1 vs. L5 were: cholesterol ester (CE) (18:2), CE (20:4), free fatty acid (FA) (20:4), phosphatidyl-choline (PC) (36:3), PC (34:2), PC (38:5), PC (36:4), PC (34:1), triacylglycerol (TG) (54:3), and PC (38:6) (all p < .001), with CE (18:2), CE (20:4), PC (36:3), PC (34:2), PC (38:5), PC (36:4), TG (54:3), and PC (38:6) independently associating with fatal events during 1-year of follow-up (all p < .05).
Conclusions
Reductions in LDL electronegativity are linked to alterations of the LDL lipidome, associate with all-cause and cardiovascular mortality beyond established risk factors, and represent a novel risk factor for adverse outcomes in patients with ACS. These associations warrant further validation in independent cohorts.
The identification of novel cardiovascular risk factors is limited by tradeoffs between statistical power, resource availability (e.g., complexity of sample preparation and analysis), and costs. Case-cohort studies have emerged as an attractive epidemiological approach to study the association between exposures and disease outcomes [
Association between skin and aortic vascular inflammation in patients with psoriasis: a case-cohort study using positron emission tomography/computed tomography.
], delineates the framework of an observational study in which a random subset of the full cohort is selected, while all newly occurring cases within the original cohort are concurrently included. This study design provides high efficiency and flexibility enabling the cost-effective investigation of multiple exposures while minimizing the risk of selection bias owing to outcome-dependent sampling, as it may occur in nested case-control studies, providing precise estimates of exposure-outcome associations of the full-cohort [
Low-density lipoprotein (LDL) indubitably promotes the initiation and progression of atherosclerotic cardiovascular disease (ASCVD), with alterations in LDL quality representing an important but understudied determinant of ASCVD risk [
Susceptibility of low-density lipoprotein particles to aggregate depends on particle lipidome, is modifiable, and associates with future cardiovascular deaths.
Sialic acid content of LDL in coronary artery disease: no evidence of desialylation in subjects with coronary stenosis and increased levels in subjects with extensive atherosclerosis and acute myocardial infarction, Arterioscler.
]. In patients with a recent acute coronary syndrome (ACS) or established ASCVD, interventions that lower levels of LDL-cholesterol (LDL-C) improve cardiovascular outcomes, but the residual risk remains high [
]. Observational studies examining the relationship between baseline LDL-C levels and mortality in patients with ACS have yielded discordant outcomes: one study has reported a counterintuitive inverse relationship [
Investigators of National Registry of Myocardial Infarction (NRMI) 4b–5, Relationship between serum low-density lipoprotein cholesterol and in-hospital mortality following acute myocardial infarction (the lipid paradox).
CCC-ACS Investigators, Revisiting the lipid paradox in ST-elevation myocardial infarction in the Chinese population: findings from the CCC-ACS project.
], potentially due to other factors influencing short-to-mid-term outcomes post-ACS. Whilst traditional cardiovascular risk factors, such as those informing the Framingham risk score [
], are undeniably linked to long-term risk of cardiovascular events, biomarkers other than cholesterol levels have been shown to determine 1-year outcomes after the index ACS [
Should patients with acute coronary disease be stratified for management according to their risk? Derivation, external validation and outcomes using the updated GRACE risk score.
LDL particles ship their water-insoluble lipid cargo in a polar shell of apolipoproteins which serve as a molecular fingerprint to direct them to specific cell types. A single apolipoprotein (apo) B100 molecule encircles the LDL particle and stabilizes the outer unilamelar layer consisting of amphiphilic phospholipids, sphingolipids and unesterified cholesterols, with its hydrophobic core containing a conglomerate of cholesteryl esters and triacylglycerols [
]. Beyond the presence of additional proteins (e.g., apoC-III) and the degree of sialylation, lipid composition represents a major determinant of LDL charge [
], the latter being particularly relevant in patients with a recent ACS. However, the association of altered LDL electronegativity and mortality in these patients is uncertain, and data on corresponding changes in the LDL particles’ lipidome remain limited.
To address this knowledge gap, we aimed (1) to determine the associations of LDL electronegativity with all-cause and cardiovascular mortality, (2) to test its predictive utility beyond and above established risk scores, and (3) to study the lipidome of least (L1) and most electronegative (L5) LDL particles in patients with ACS who were prospectively recruited at four university hospitals in Switzerland.
2. Patients and methods
2.1 Study design and participants
This is a case-cohort study nested within the prospective, multicentre SPUM-ACS study. The design of case-cohort studies [
Association between skin and aortic vascular inflammation in patients with psoriasis: a case-cohort study using positron emission tomography/computed tomography.
Sex-specific evaluation and redevelopment of the GRACE score in non-ST-segment elevation acute coronary syndromes in populations from the UK and Switzerland: a multinational analysis with external cohort validation.
]. Briefly, from October 2012 until December 2017 a total of 2619 patients with ACS aged ≥18 years presenting within 5 days after pain onset were recruited at four university hospitals in Switzerland (SPUM-ACS study; Cohort II; ClinicalTrials.gov Identifier: NCT01000701). EDTA-plasma samples were obtained prior to coronary angiography. Data on baseline demographics, risk factors, and medication were entered by trained personnel using a centralized data entry system. All patients were followed at 30 days and 1 year. The mortality endpoint was reviewed by an independent event adjudication committee comprising three certified cardiologists using predefined adjudication forms and blinded to patients’ baseline characteristics. Of all 1272 patients with follow-up data and ≥2.0 ml EDTA-plasma for fast protein liquid chromatography (FPLC) available, a subcohort was drawn using simple random sampling. The oversampling of cases inherent to the case-cohort design was accounted for by fitting weighted Cox proportional hazard regression models (Supplemental Material) [
]. All participants provided informed consent. The study was conducted according to the Declaration of Helsinki, and was approved by the institutional review board. We followed the principles outlined by the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) initiative for the reporting of cohort studies. The STROBE checklist and details on case-cohort sampling methods are provided in the Supplemental Material.
2.2 Assessment of LDL-cholesterol, electronegativity and lipidome
Informed by standard lipid panels, levels of low-density lipoprotein cholesterol (LDL-C) were calculated using the Sampson equation [
]. Isolation of LDL was achieved by sequential KBr-based ultracentrifugation (1.019–1.063 g/mL), and dissolved into its 5 subfractions with increasing electronegativity (L1 to L5) using anion-exchange columns on a fast protein liquid chromatography (FPLC) system [
]. The effluent, yielded by a multistep sodium chloride gradient, was monitored at 280 nm, and concentrations of L1 and L5 (least and most electronegative subfraction, respectively) were calculated, as previously described [
], with the L1-L5 ratio serving as a proxy of overall LDL electronegativity. Least and most electronegative LDL particles (L1 and L5, respectively) of a subcohort of patients were subjected to untargeted lipidomics, as previously reported (Supplemental Material) [
]. Only lipid species with known annotation (according to LIPID MAPS® Structure Database; accessed on March 22, 2022; Supplemental Table 1) were considered for statistical analyses [
]. LDL electronegativity and lipidomics data were generated by blinded study personnel.
2.3 Statistical analysis
Continuous data are presented as mean and standard deviations (SD) or median and interquartile ranges (IQRs), and categorical variables as counts and percentages (%). Given the oversampling of cases, covariate balance was assessed by calculating standardized mean differences (SMD). We estimated crude (HR) and adjusted hazard ratios (aHR) with 95% confidence intervals (CIs) for 30-day and 1-year mortality per SD increment of each biomarker in the full cohort by fitting weighted Cox proportional hazard regression models [
]. Model 1 represents the crude regression model, model 2 includes sex (categorical) and age (continuous), and model 3 includes sex (categorical), age (continuous), high-sensitivity cardiac troponin T (hs-cTnT; continuous), diabetes (categorical), estimated glomerular filtration rate (eGFR; continuous), statin use (categorical), and LDL-C levels (continuous), as specified in the figure legend. Model 4 additionally accounts for GRACE 2.0 risk for 1-year death (<3% low, 3–8% intermediate and >8% high risk) [
]. In sensitivity analyses, all models were additionally adjusted for the time elapsed between symptom onset and blood sampling (continuous) (Supplemental Material). Where data on covariates in the respective regression models were missing (Supplemental Table 2), we applied multiple imputation by chained equations (MICE; Supplemental Tables 3–5). Independent variables included in the multivariable-adjusted regression models were ranked based on their Wald χ2 value, as reported (Supplemental Material) [
]. Potential effects of MICE on main results were explored in additional sensitivity analyses using complete cases (Supplemental Tables 5 and 7). To assess whether the L1-L5 ratio improves the performance of the GRACE 2.0 or Framingham risk score to predict 1-year death while accounting for different sampling weights, time-dependent area under the receiver operating characteristic curve (AUC) and goodness of fit (likelihood function) of both risk prediction models (i.e., GRACE 2.0 or the Framingham and the biomarker-enhanced GRACE 2.0 or Framingham model) were calculated following stratified superpopulation sampling [
Should patients with acute coronary disease be stratified for management according to their risk? Derivation, external validation and outcomes using the updated GRACE risk score.
]. Additional analyses were focused on the associations of top 10 lipid species enriched in least (L1) electronegative LDL particles with outcomes of interest by fitting conditional logistic regression models, further adjusting for sex (categorical) and age (continuous) (Supplemental Tables 7–8). Statistical significance for all analyses was established at p < .05. All analyses were performed in R version 4.2.1. A detailed description is provided in the Supplemental Material.
3. Results
3.1 Study population
The present case-cohort study is based on data from a subset of 2619 patients with ACS recruited between October 17, 2012 and December 31, 2017. Among the final study population comprising 1272 patients a total of 52 deaths (4%) occurred within 1 year of follow-up (Fig. 1). At baseline, cases differed in several clinical features associated with increased cardiovascular risk as compared to patients sampled in the subcohort (Table 1). Patients who had a fatal event during follow-up presented at a higher age (median [IQR], 76.45 (67.67–82.22) and 73.95 (65.58–80.97) years; SMD, 0.22), with lower estimated glomerular filtration rate (eGFR; 62.16 (44.66–78.86) and 78.15 (59.26–86.96) ml/min/1.73 m2, 0.55) and were more often diabetics (n (%), 22 (42) and 45 (30), 0.26). Moreover, they showed higher LDL-C levels (102.03 (72.52–123.19) and 87.66 (67.05–119.46) mg/dl, 0.19), were more frequently assigned to the GRACE high-risk group (37 (82) and 70 (54), 0.75) and had a noSupplemental Table hift in the electronegative properties of isolated LDL particles, implying the presence of less negatively charged LDL (L1-L5 ratio, 4.79 (2.80–8.55) and 3.83 (2.03–7.73), 0.236).
Fig. 1Flow diagram of the SPUM-ACS case-cohort study. aDefined as grossly haemolytic or plasma volume <2.0 ml. bPatients who failed to complete the study such as those who withdrew consent or were lost to follow-up were censored at the date of last contact or the date of the assessment of the survival status, whichever occurred later.
Defined as <3% (low-risk), 3–8% (intermediate-risk), or >8% (high-risk).
0.657
0.745
Low-risk
15 (11.5)
1 (2.2)
1 (2.4)
Intermediate-risk
45 (34.6)
7 (15.6)
5 (11.9)
High-risk
70 (53.8)
37 (82.2)
36 (85.7)
Continuous data are mean (SD) or median (IQR) if skewed and categorical data are n (%). Abbreviations: ACEi, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; GRACE 2.0, Global Registry of Acute Coronary Events; HDL-C, high-density lipoprotein cholesterol; hs-cTnT, high-sensitivity cardiac troponin; LDL-C, low-density lipoprotein cholesterol, SMD standardized mean difference. SI conversion factors: To convert cholesterol levels to millimoles per liter, multiply by 0.0259.
The number of events occurring within/outside the subcohort are specified in Fig. 1.
a Includes cases within and outside the subcohort and study participants may have had multiple incident events.
b Percentages may not total 100 owing to rounding.
c Defined as <3% (low-risk), 3–8% (intermediate-risk), or >8% (high-risk).
3.2 Independent association of altered LDL electronegativity and fatal outcomes
In unadjusted analyses, reductions in overall LDL electronegativity, as assessed by the L1-L5 ratio, were strongly linked to mortality from any and cardiovascular causes at both 30 days (hazard ratio (HR), 2.18, 95% CI, 1.28–3.69, p=.0041; and 2.31, 1.37–3.90 per 1 SD increment, p=.0020) and 1 year (aHR, 1.91, 1.20–3.04, p=.0065; and 1.89, 1.19–3.02 per 1 SD increment, p=.0077), which remained consistent in multivariable-adjusted analyses accounting for established clinical risk factors (adjusted [a]HR, 2.13, 1.07–4.23, p=.032; and 2.29, 1.21–4.35, p=.011; 1.84, 1.03–3.29, p=.038; and 1.88, 1.08–3.28, p=.027; Fig. 2A and B; Supplemental Table 3). These associations were robust in sensitivity analyses considering an alternative set of potential confounders, including the time elapsed between symptom onset and blood sampling (aHR, 2.18, 1.09–4.39, p=.029; 1.86, 1.05–3.31, p=.033; 2.34, 1.19–4.60, p=.015; 1.89, 1.07–3.33, p=.028; Supplemental Table 4). Complete-case analyses yielded similar results (Supplemental Table 5). Conversely, no associations between baseline LDL-C levels and all-cause or cardiovascular mortality were observed at 30 days or 1 year (Supplemental Fig. 1; Supplemental Tables 6–7).
Fig. 2Associations of changed LDL electronegativity with mortality.
Data are HRs and 95% confidence intervals of death from any (A) or cardiovascular causes (B) at 30 days and 1 year per SD increment in L1/L5. Model 1 represents the crude regression model; model 2 includes sex (categorical) and age (continuous); model 3 includes sex (categorical), age (continuous), high-sensitivity cardiac troponin T (continuous), diabetes (categorical), estimated glomerular filtration rate (continuous), statin use (categorical), and levels of LDL-C (continuous); model 4 includes sex (categorical), age (continuous), high-sensitivity cardiac troponin T (continuous), diabetes (categorical), estimated glomerular filtration rate (continuous), statin use (categorical), levels of LDL-C (continuous), and GRACE 2.0 1-year mortality risk estimates (categorical).
3.3 LDL electronegativity and 1-year mortality risk beyond and above the updated GRACE or Framingham risk score
Among all variables included in the multivariable-adjusted regression model to predict 1-year outcomes, including established clinical risk factors, such as LDL-C, the L1-L5 ratio was the second highest-ranked predictor of death from any cause (Fig. 3A), and superseded several clinical risk factors for the prediction of cardiovascular death (Fig. 3B). Notably, changes in LDL electronegativity associated with 1-year mortality risk beyond the one estimated by GRACE 2.0 (aHR, 1.92, 1.08–3.43, p=.027), suggesting a potential predictive utility of this biomarker over and above the updated GRACE risk score. Indeed, adding the L1-L5 ratio to the GRACE model increased its goodness of fit (p < .001) and resulted in improved discriminatory performance as compared to the original GRACE 2.0 risk score (AUC, 0.78 vs. 0.74, p=.03; Fig. 4A). Conversely, adding the L1-L5 ratio to the Framingham risk score did not improve its discriminatory performance (AUC, 0.59 vs. 0.62, p=.83; Fig. 4B).
Fig. 3Relative effect of each independent variable on model output.
Fig. 4Smoothed receiver operating characteristic curve of the GRACE 2.0 and biomarker (L1-L5 ratio)-enhanced GRACE 2.0 risk score (A) or Framingham and biomarker-enhanced Framingham risk score (B) for the prediction of 1-year death. Paired ROC curves were compared by bootstrapping using 5000 replicates. GRACE denotes global registry of acute coronary events.
3.4 Lipidome of LDL particles differs according to their electronegative properties and determines mortality risk
Lipidomic analyses of least (L1) and most (L5) electronegative LDL particles unveiled abundancy of fatty acyls, glycerolipids, glycerophospholipids, sphingolipids, and sterol lipids (Fig. 5; Supplemental Table 1). The top 10 differentially expressed lipid species, ranked by the mean difference between least (L1) and most (L5) electronegative LDL particles, were: cholesterol ester (CE) (18:2), CE (20:4), free fatty acid (FA) (20:4), phosphatidylcholine (PC) (36:3), PC (34:2), PC (38:5), PC (36:4), PC (34:1), triacylglycerol (TG) (54:3), and PC (38:6) (all p < .001; Supplemental Table 8). In exploratory analyses, 8 out of these 10 lipid species enriched in L1 LDL showed an independent association with mortality from any and cardiovascular causes at 1 year (CE (18:2), adjusted odds ratio (aOR), 4.23 and 3.99; CE (20:4), 6.21 and 5.44; PC (36:3), 5.32 and 4.86; PC (34:2), 4.76 and 4.36; PC (38:5), 9.05 and 7.63; PC (36:4), 5.34 and 4.75; TG (54:3), 4.66 and 4.25; PC (38:6), 5.48 and 5.03 per 1 SD increment in each lipid species; p < .05; Supplemental Table 9).
Fig. 5Customized Manhattan plot for the comparative analysis of 482 annotated lipid species of least (L1) vs. most electronegative (L5) LDL particles.
Individual lipid species arranged by 25 lipid classes (y-axis) and -log10 P values (x-axis) for their association with L1 vs. L5 LDL particles. The threshold of the false discovery rate of less than 0.05 is signified by the vertical dashed line. Color indicates the magnitude of mean difference in normalized expression with triangularly shaped dots representing the 10 highest-ranked lipid species. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)
In a prospective cohort of contemporary patients with ACS, we demonstrate for the first time that attenuated LDL electronegativity, as assessed by the L1-L5 ratio, unfavorably associates with all-cause and cardiovascular mortality after the index ACS, over and above the updated GRACE risk score. Moreover, our data suggest that the lipidome of LDL particles markedly differs depending on their charge, with the majority of highest-ranked lipid species independently associating with adverse events. Importantly, LDL electronegativity represented the second highest-ranked predictor of all-cause death (Fig. 3A), and superseded several clinical risk factors for the prediction of cardiovascular death at 1 year (Fig. 3B). While the exact mechanism underpinning this phenomenon warrants further study, it is interesting to note that electronegative properties of LDL particles determine their LDL-receptor (LDLR) affinity, with changes in LDL charge impinging on their pro-thrombotic and pro-atherogenic effects [
]. In fact, while PCSK9 inhibition on top of statin-therapy associates with a marked risk reduction in myocardial infarction, ischemic stroke, and coronary revascularization, the short-to-mid-term mortality benefit of aggressive LDL-C lowering appears marginal [
]. In the present study, baseline LDL-C levels showed no association with all-cause or cardiovascular mortality at 1 year after the index ACS, yet hazard ratios nominally increased across follow-up periods within each model (Supplemental Tables 6–7; Supplemental Fig. 1), suggesting a potential association at long-term follow-up. Whilst interleukin-6 (IL6)-driven LDL-receptor upregulation may attenuate a potential association in the acute setting [
], results remained consistent after controlling for high-sensitivity cardiac troponin T (hs-cTnT), the latter showing release kinetics similar to those of IL6 following acute myocardial ischaemia [
Importantly, the L1-L5 ratio provided additive predictive utility beyond the updated GRACE (Fig. 4A) but not Framingham risk score, emphasizing the importance of this biomarker for the prediction of adverse events in patients with ACS. Indeed, while the Framingham risk score shows good discriminatory performance for the prediction of 10-year cardiovascular disease (CVD) events (defined as a composite of clinical events related to coronary heart disease, cerebrovascular disease, peripheral artery disease, and heart failure) in individuals free of CVD [
], as similarly observed in magnitude in the present analysis focusing on 1-year outcomes (Fig. 4B). This might be due to several factors, including the different study population (healthy individuals vs. patients with ACS), distinct endpoints (10-year CVD events vs. 1-year mortality), and different study settings (primary vs. secondary prevention). As such, the GRACE risk score, derived and validated in patients with ACS [
Should patients with acute coronary disease be stratified for management according to their risk? Derivation, external validation and outcomes using the updated GRACE risk score.
Susceptibility of low-density lipoprotein particles to aggregate depends on particle lipidome, is modifiable, and associates with future cardiovascular deaths.
], a patient population at particularly high cardiovascular risk. In fact, lipidomic analyses revealed an enrichment in atherogenic lipid species in least (L1) electronegative lipid particles (Supplemental Table 8), with the majority thereof independently associating with risk of mortality from any and cardiovascular causes (Supplemental Table 9). For instance, L1 was found to be highly enriched in cholesterol ester (CE) (18:2) and CE (20:4), two lipid species abundantly expressed in atherosclerotic plaques [
Susceptibility of low-density lipoprotein particles to aggregate depends on particle lipidome, is modifiable, and associates with future cardiovascular deaths.
], thereby providing a potential soil for progressive ASCVD.
To the best of our knowledge, this is the first cohort study examining the relationship of LDL electronegativity with adverse outcomes in patients with established ASCVD. The independent association of the L1-L5 ratio with multiple endpoints using different regression models that account for a variety of potential confounders, and its additive predictive value beyond an established risk prediction model provide high internal validity of our findings (Supplemental Material). Nonetheless, external validation studies are warranted to confirm the independent association of LDL electronegativity with mortality in patients with ACS.
4.1 Conclusions
In contemporary patients with ACS, a reduction in overall LDL electronegativity associates with all-cause and cardiovascular mortality beyond established risk factors, including LDL-C, and provides additional prognostic utility beyond the updated GRACE score (Fig. 6). Electronegative properties of LDL particles are tightly linked to alterations in their lipidome and represent a novel determinant of mortality risk in patients with ACS. These findings should stimulate further research into LDL quality as a potential risk factor of ASCVD initiation and accelerated disease progression. Ultimately, these efforts may open novel therapeutic avenues that go beyond LDL-C lowering for secondary prevention of ASCVD.
Fig. 6In prospectively recruited patients with acute coronary syndromes (ACS), a reduction in overall low-density lipoprotein (LDL) electronegativity emerged as a strong predictor of all-cause and cardiovascular mortality independent of established risk factors, including LDL-C, and provided additional prognostic utility beyond the updated GRACE risk score.
Changes in the electronegative properties of LDL particles are tightly linked to alterations in their lipidome and represent a novel determinant of mortality risk in patients with established atherosclerotic cardiovascular disease (ASCVD). These observations should trigger further research into LDL quality as a potentially understudied yet important risk factor of ASCVD initiation and accelerated disease progression. FPLC denotes fast protein liquid chromatography, and FU follow-up.
To the best of our knowledge, this is the first cohort study on LDL electronegativity in patients with ACS. Our study has several strengths, including that it is based on a large, multicentre, prospective cohort of contemporary patients with ACS, with robust follow-up data on prespecified end points, and independent clinical event adjudication [
Sex-specific evaluation and redevelopment of the GRACE score in non-ST-segment elevation acute coronary syndromes in populations from the UK and Switzerland: a multinational analysis with external cohort validation.
]. By applying a case-cohort design, we could optimize study resources while preserving the benefits of cohort studies to make patient population-based inferences [
]. The present study has several limitations inherent to any observational study, including residual confounding. Furthermore, although regression models accounting for the time elapsed between symptom onset and biomarker measurements yielded consistent results, future studies are warranted to assess whether changes in LDL charge occur acutely and are also determined by prehospital delays. Indeed, experimental data on why and how shifts in LDL charge occur are scarce; thus, preclinical efforts disentangling the mechanistic basis of altered LDL electronegativity should be continued incessantly, aiming toward the future goal to assess clinical applicability of this novel marker, including possible therapeutic avenues. Finally, the generalizability of our results to patients of other ethnicities and/or stable coronary disease warrants further study.
Financial support
This study was supported by funding granted by the Swiss National Science Foundation (SPUM 33CM30-124112 and 32473B_163271; to TFL) and the Swiss Heart Foundation (to TFL). Additional funding of the SPUM-ACS study was received by unrestricted grants from Roche Diagnostics (Boehringer Mannheim; Indianapolis, IN, USA), Eli Lilly (Indianapolis, IN, USA), AstraZeneca (Baar, Switzerland), Medtronic (Münchenbuchsee, Switzerland), Merck Sharpe and Dohme (Lucerne, Switzerland), Sanofi-Aventis (Vernier, Switzerland), and St Jude Medical AG (Zurich, Switzerland). This work was further supported by the Foundation for Cardiovascular Research - Zurich Heart House (to SK, FAW, GGC, AA, and TFL), the Lindenhofstiftung (to TFL USA and SK), and the Theodor Ida Herzog-Egli Stiftung (to SK). The funding sources had no role in study design, data collection, data analysis, data interpretation, or writing of the report.
Data availability statement
The data underlying this study will be made available to other researchers upon reasonable request to the corresponding authors, subject to institutional and ethical committee approvals.
Clinical trial number
ClinicalTrials.gov Identifier: NCT01000701.
CRediT authorship contribution statement
Simon Kraler: had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis, Conceptualization, Study concept and design, Funding acquisition, Acquisition, analysis, Formal analysis, or interpretation of data, Writing – original draft, Drafting of the manuscript, Statistical analysis, Obtained funding, Project administration, Administrative, technical, or material support, Critical revision of the manuscript. Florian A. Wenzl: Funding acquisition, Acquisition, analysis, or interpretation of data, Formal analysis. Jody Vykoukal: Funding acquisition, Acquisition, analysis, or interpretation of data, Formal analysis. Johannes F. Fahrmann: Funding acquisition, Acquisition, analysis, or interpretation of data, Formal analysis. Ming-Yi Shen: Funding acquisition, Acquisition, analysis, or interpretation of data, Formal analysis. Der-Yuan Chen: Funding acquisition, Acquisition, analysis, or interpretation of data, Formal analysis. Kuan-Cheng Chang: Funding acquisition, Acquisition, analysis, or interpretation of data, Formal analysis. Ching-Kun Chang: Funding acquisition, Acquisition, analysis, or interpretation of data, Formal analysis. Arnold von Eckardstein: Funding acquisition, Acquisition, analysis, or interpretation of data, Formal analysis. Lorenz Räber: Administrative, Project administration, technical, or material support. François Mach: Administrative, Project administration, technical, or material support. David Nanchen: Administrative, Project administration, technical, or material support. Christian M. Matter: Administrative, Project administration, technical, or material support. Luca Liberale: Administrative, Project administration, technical, or material support, Funding acquisition, Acquisition, analysis, or interpretation of data, Formal analysis. Giovanni G. Camici: Administrative, Project administration, technical, or material support. Alexander Akhmedov: Funding acquisition, Acquisition, analysis, Formal analysis, or interpretation of data, Conceptualization, Study concept and design, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis, Supervision, Study supervision. Chu-Huang Chen: Funding acquisition, Acquisition, analysis, Formal analysis, or interpretation of data, Conceptualization, Study concept and design, Project administration, Administrative, technical, or material support, Supervision, Study supervision. Thomas F. Lüscher: Funding acquisition, Acquisition, analysis, Formal analysis, or interpretation of data, Writing – original draft, Drafting of the manuscript, Conceptualization, Study concept and design, Obtained funding, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis, Project administration, Administrative, technical, or material support, Supervision, Study supervision, Critical revision of the manuscript.
Declaration of competing interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: SK received funding from the Swiss Heart Foundation, the Lindenhof Foundation, the Novartis Foundation for Medical-biological Research, and the Theodor-Ida-Herzog-Egli Foundation, and equipment and materials from Roche Diagnostics outside the submitted work. LR received funding from Abbott, Biotronik, Boston Scientific, Heartflow, Sanofi, and Regeneron, and declares consulting fees from Abbott, Amgen, AstraZeneca, Canon, Medtronic, NovoNordisk, Occlutech, Sanofi, and Vifor, payment or honoraria from Abbott and Occlutech, and travel support from AstraZeneca. FM has received research grants to the institution from Amgen, AstraZeneca, Boston Scientific, Biotronik, Eli Lilly, Medtronic, MSD, and St. Jude Medical, including speaker and/or consultant fees. AvE received speaker and/or consultant fees from Amgen, MSD, and Sanofi-Aventis. CMM received research grants to the institution from Eli Lilly, AstraZeneca, Roche, Amgen and MSD including speaker or consultant fees. G.G.C. and L.L. are co-inventors on the international patent WO/2020/226993 filed in April 2020. The patent relates to the use of antibodies which specifically bind IL-1α to reduce various sequelae of ischaemia–reperfusion injury to the central nervous system. G.G.C. is a consultant to Sovida Solutions Limited. L.L. reports speaker fees from Daiichi-Sankyo outside the submitted work. TFL declares institutional educational and research grants from Abbott, Amgen, AstraZeneca, Boehringer Ingelheim, Daiichi Sankyo, Novartis, and Vifor, and consulting fees from Daiichi Sankyo, Ineeo Inc, Philipps, and Pfizer outside the submitted work. TFL holds leadership positions at the European Society of Cardiology, Swiss Heart Foundation, and the Foundation for Cardiovascular Research—Zurich Heart House. All other authors declare no competing interests.
Acknowledgements
We are grateful to all health-care professionals who participated in the conduct of the national registry this study is based on. We are particularly grateful to the members of the independent events adjudication committee (Matthias Pfisterer, Lukas Kappenberger, Tiziano Moccetti, David Carballo, Baris Gencer, Philippe Meyer). The authors are also grateful to members of the Clinical Trial Unit, University of Bern, Bern, Switzerland for handling the database. We further express our gratitude towards all administrative and IT personnel involved in data entry and management.
Appendix A. Supplementary data
The following is the Supplementary data to this article.
Associations of LDL-C levels with mortality from any (top) and cardiovascular causes (bottom). Data are HRs and 95% confidence intervals of death from any or cardiovascular causes at 30 days and 1 year per SD increment in baseline levels of LDL-C. Model 1 is the crude regression model; model 2 includes sex and age (continuous); model 3 includes sex, age (continuous), high-sensitivity cardiac troponin T (continuous), diabetes (categorical), estimated glomerular filtration rate (continuous), and statin use (categorical); model 4 includes sex (categorical), age (continuous), high-sensitivity cardiac troponin T (continuous), diabetes (categorical), estimated glomerular filtration rate (continuous), statin use (categorical), and GRACE 2.0 1-year mortality risk estimates (categorical).
References
Narula S.
Yusuf S.
Chong M.
Ramasundarahettige C.
Rangarajan S.
Bangdiwala S.I.
van Eikels M.
Leineweber K.
Wu A.
Pigeyre M.
Paré G.
Plasma ACE2 and risk of death or cardiometabolic diseases: a case-cohort analysis.
Association between skin and aortic vascular inflammation in patients with psoriasis: a case-cohort study using positron emission tomography/computed tomography.
Susceptibility of low-density lipoprotein particles to aggregate depends on particle lipidome, is modifiable, and associates with future cardiovascular deaths.
Sialic acid content of LDL in coronary artery disease: no evidence of desialylation in subjects with coronary stenosis and increased levels in subjects with extensive atherosclerosis and acute myocardial infarction, Arterioscler.
Investigators of National Registry of Myocardial Infarction (NRMI) 4b–5, Relationship between serum low-density lipoprotein cholesterol and in-hospital mortality following acute myocardial infarction (the lipid paradox).
CCC-ACS Investigators, Revisiting the lipid paradox in ST-elevation myocardial infarction in the Chinese population: findings from the CCC-ACS project.
Should patients with acute coronary disease be stratified for management according to their risk? Derivation, external validation and outcomes using the updated GRACE risk score.
Sex-specific evaluation and redevelopment of the GRACE score in non-ST-segment elevation acute coronary syndromes in populations from the UK and Switzerland: a multinational analysis with external cohort validation.
00207-1&id=ga1.jpg){kind=link}
00207-1&id=gr1.jpg){kind=link}
00207-1&id=gr2.jpg){kind=link}
00207-1&id=gr3.jpg){kind=link}
00207-1&id=gr4.jpg){kind=link}
00207-1&id=gr5.jpg){kind=link}
00207-1&id=gr6.jpg){kind=link}