Acute Coronary Syndrome Subphenotypes Based on Repeated Biomarker Measurements in Relation to Long‐Term Mortality Risk
Abstract
Background
We aimed to identify patients with subphenotypes of postacute coronary syndrome (ACS) using repeated measurements of high‐sensitivity cardiac troponin T, N‐terminal pro‐B‐type natriuretic peptide, high‐sensitivity C‐reactive protein, and growth differentiation factor 15 in the year after the index admission, and to investigate their association with long‐term mortality risk.
Methods and Results
BIOMArCS (BIOMarker Study to Identify the Acute Risk of a Coronary Syndrome) was an observational study of patients with ACS, who underwent high‐frequency blood sampling for 1 year. Biomarkers were measured in a median of 16 repeated samples per individual. Cluster analysis was performed to identify biomarker‐based subphenotypes in 723 patients without a repeat ACS in the first year. Patients with a repeat ACS (N=36) were considered a separate cluster. Differences in all‐cause death were evaluated using accelerated failure time models (median follow‐up, 9.1 years; 141 deaths). Three biomarker‐based clusters were identified: cluster 1 showed low and stable biomarker concentrations, cluster 2 had elevated concentrations that subsequently decreased, and cluster 3 showed persistently elevated concentrations. The temporal biomarker patterns of patients in cluster 3 were similar to those with a repeat ACS during the first year. Clusters 1 and 2 had a similar and favorable long‐term mortality risk. Cluster 3 had the highest mortality risk. The adjusted survival time ratio was 0.64 (95% CI, 0.44–0.93; P=0.018) compared with cluster 1, and 0.71 (95% CI, 0.39–1.32; P=0.281) compared with patients with a repeat ACS.
Conclusions
Patients with subphenotypes of post‐ACS with different all‐cause mortality risks during long‐term follow‐up can be identified on the basis of repeatedly measured cardiovascular biomarkers. Patients with persistently elevated biomarkers have the worst outcomes, regardless of whether they experienced a repeat ACS in the first year.
Nonstandard Abbreviations and Acronyms
- BIOMArCS
- Biomarker Study to Identify the Acute Risk of a Coronary Syndrome
- GDF‐15
- growth differentiation factor 15
- hs‐cTnT
- high‐sensitivity cardiac troponin T
Clinical Perspective
What Is New?
•
We report a comprehensive cluster analysis of repeatedly measured high‐sensitivity troponin T, N‐terminal pro‐B‐type natriuretic peptide, high‐sensitivity C‐reactive protein, and growth differentiation factor 15 in the year following an acute coronary syndrome admission, and investigated their association with long‐term mortality risk.
•
In the low‐risk patients who had no repeat acute coronary syndrome in the first year, 3 subphenotypes were identified, displaying significant differences in their longitudinal biomarker profile: cluster 1 showed relatively low and stable biomarker concentrations, cluster 2 had elevated concentrations that subsequently decreased, and cluster 3 showed persistently elevated concentrations.
•
The low‐risk patients with persistently elevated biomarker concentrations (cluster 3) had adverse prognoses similar to those who experienced a repeat acute coronary syndrome in the first year. Patients with relatively low biomarker concentrations at 1 year follow‐up (clusters 1 and 2) showed the best prognosis.
What Are the Clinical Implications?
•
Our findings highlight the importance of biomarkers for long‐term acute coronary syndorme prognostication and personalized risk assessment.
•
Incorporating both favorable and unfavorable biomarker profiles into clinical practice could potentially foster targeted care to individuals who may require different levels of monitoring and intervention.
In recent decades, circulating biomarkers such as cardiac troponin and NT‐proBNP (N‐terminal pro‐B‐type natriuretic peptide) have been shown to provide objective and accurate prognostic information in patients with established coronary artery disease.1, 2, 3, 4, 5 These cardiac biomarkers could facilitate tailoring of appropriate treatment in individual patients.6, 7, 8, 9 Obviously, while pursuing personalized management, appropriate longitudinal risk assessment and stratification of patients are crucial. In this context, clustering patients with established coronary artery disease based on dynamic changes in their cardiovascular biomarker profile may be a powerful tool to provide information on individual patients' risk and to improve prognostication. Nonetheless, studies clustering patients with cardiac disease based on their biomarker profile are scarce, mostly performed in patient cohorts with heart failure,10, 11, 12 and generally limited to biomarker measurements at 1 point in time.13, 14
BIOMArCS (the Biomarker Study to Identify the Acute Risk of a Coronary Syndrome) was specifically designed to study the longitudinal biomarker patterns in patients admitted for an acute coronary syndrome (ACS).15, 16 BIOMArCS enrolled 844 patients, who had a median of 17 repeated blood samples taken in the first year after the index event. Concentrations of hs‐cTnT (high‐sensitivity cardiac troponin T), NT‐proBNP, hs‐CRP (high‐sensitivity C‐reactive protein), and GDF‐15 (growth differentiation factor 15) were measured in these samples, and a comprehensive unsupervised clustering analysis was conducted to identify biomarker‐based phenotypes. The clinical characteristics of the resulting clusters were then studied, along with their relationship to all‐cause death during long‐term follow‐up. Of particular interest were the dynamic biomarker patterns of low‐risk patients who did not experience a repeat ACS in the first year. Patients who did experience a repeat ACS are considered high‐risk patients17 and are therefore, for comparison, studied as a separate cluster.
METHODS
Data Availability Statement
The data that support the findings of this study are available from the corresponding author upon reasonable request.
Study Population
BIOMArCS is an observational study that was conducted in 18 hospitals across the Netherlands from 2008 to 2015.15, 16 The study aimed to collect data on biomarker patterns in patients with ACS during a 1‐year follow‐up period. BIOMArCS enrolled patients aged ≥40 years who were admitted with an ACS and had at least 1 cardiovascular risk factor. Patients with severely impaired ejection fraction, severe chronic kidney disease, or a coexisting condition with a life expectancy of <1 year were excluded. Patients underwent regular blood sampling in the first year after the index ACS admission according to a strict schedule.16 Ultimately, a total of 12 218 blood samples were obtained from 844 patients, with a median of 17 repeated blood samples per patient. The patients received treatment according to prevailing guidelines and at the discretion of the treating physician. The current study included the 723 patients who were known to be alive 1 year after the index event, had at least 1 biomarker measurement available and did not experience a repeat ACS during the first year after the index event. Additionally, 36 patients who did experience a nonfatal repeat ACS were also included (see Figure S1 for details on patient selection).
BIOMArCS was conducted in compliance with the Declaration of Helsinki, and the study protocol was approved by the institutional review boards of the participating hospitals. All study subjects provided written informed consent. BIOMArCS is registered in The Netherlands Trial Register with the unique identifiers NTR1106 and NTR1698.
Sample Collection and Processing
To ensure the quality and reliability of the data, standardized protocols for sample collection, handling, and storage were applied. Blood samples were processed and stored at −80°C within a median of 82 minutes (25th to 75th percentile, 58–117) after collection.15 The aliquots were then transported to the Department of Clinical Chemistry, Erasmus MC, Rotterdam, for long‐term storage and batchwise central analysis. Storage duration was not correlated with the included biomarkers (Figure S2). To conduct the current investigation, specific assays were used to analyze the concentrations of hs‐cTnT, NT‐proBNP, GDF‐15, and hs‐CRP. The hs‐cTnT, NT‐proBNP, and GDF‐15 concentrations were analyzed in a single batch using Elecsys quantitative sandwich electro‐chemiluminescence immunoassays ECLIA on a cobas e 601 analyzer (Roche Diagnostics, Ltd., Rotkreuz, Switzerland). Meanwhile, hs‐CRP concentrations were analyzed using a particle‐enhanced immunoturbidimetric assay on a cobas c 501 analyzer (Roche Diagnostics). The study used 10 606 samples, with a median of 16 samples per patient, that were obtained in the period >30 days after the index ACS, during which patients were clinically stabilized. Samples collected within the 0‐ to 30‐day window after a repeat ACS were excluded. Additional details on sample selection can be found in Figure S1.
Primary End Point
In the present study, the primary end point was all‐cause death after the 1‐year landmark point, which aligns with the end of the protocolized high‐frequency blood sampling schedule of the BIOMArCS study. Vital status data were obtained from municipal registries, and follow‐up on all‐cause death was completed until January 2021 for 96.8% of patients. Patients with incomplete follow‐up data were censored at the date they were last known to be alive.
Statistical Analysis
Biomarker measurements were studied as continuous variables, and values below the limit of blank (for hs‐cTnT) or limit of detection (for NT‐proBNP, hs‐CRP, and GDF‐15) were assigned a value at the limit of blank or limit of detection value divided by 2. A log2 transformation was applied to reduce skewness. Outliers in the data were defined as values with a distance >3 SDs from the mean and were excluded from analyses. Accordingly, 13 high‐sensitivity troponin T measurements were excluded from analyses.
A cluster analysis was performed on the repeated biomarker data of the 723 patients who remained event free during the first year following the index ACS. The first step involved the application of linear mixed‐effects modeling to estimate the concentrations of individual patients' biomarker concentrations at 30 days after the index ACS (ie, intercept coefficient), as well as the rate of change during the 30‐day to 1‐year post‐ACS period (ie, slope coefficient). Specifically, both a random intercept and slope were modeled as random effects in the linear mixed‐effects models (Data S1), using sampling time during follow‐up as the time scale for slope. The intercept and slope coefficients were then subjected to 98% winsorization and standardization. Subsequently, K‐means cluster analysis was performed on these model parameters, representing the complete trajectory of the biomarkers. K‐means is an unsupervised clustering approach. The “optimal” number of clusters was determined using the NbClust R package,18, 19 which uses the majority rule of 30 separate indices. Cluster stability was verified by resampling using several schemes, including bootstrapping and noise replacement. Clusters with a Jaccard means >0.75 were considered valid and stable.20, 21 Following this methodology, the optimal number of biomarker‐trajectory‐based clusters in the patients who did not experience a repeat ACS turned out to be 3 (see Results section). Therefore, an unsupervised K‐means cluster analysis was finally applied to partition these patients into 3 clusters.22 The patients who experienced a repeat ACS in the first year after the index event were considered a separate fourth cluster.
The clinical characteristics of the patients in the identified clusters were compared using ANOVA, Kruskal‐Wallis tests, χ2 tests, and Fisher's exact tests, depending on variable distributions. To study all‐cause death after the 1‐year landmark, the Kaplan–Meier method was used, and differences between clusters were evaluated by Fleming–Harrington statistics with ρ=1, γ=1, which is appropriate for crossing survival curves.23 Additionally, due to the observed violation of the proportional hazard assumption, accelerated failure time models with a Weibull distribution of the error terms were fitted. The survival time ratio produced by accelerated failure time models can be interpreted as the average survival time since the 1‐year landmark of 1 cluster relative to a reference cluster. Unadjusted and adjusted accelerated failure time models were explored, where adjustment factors included age, sex, diabetes, hypertension, hypercholesterolemia, smoking status, body mass index, history of coronary artery disease, history of stroke, and estimated glomerular filtration rate.
Single imputation was used to account for missing values of the clinical data that were included in the survival models using the mice package.24 All statistical analyses were conducted using R software version 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria). Statistical significance was defined as a 2‐sided P value of <0.05.
RESULTS
Clustering Based on Repeated Biomarker Measurements
Three distinct clusters were identified on the basis of the longitudinal biomarker profile of patients who did not experience a repeat ACS during the first year (see Figure 1 and Table 1). In cluster 1272 (38%) patients had relatively low and stable biomarker concentrations during the 30‐day to 1‐year post‐ACS period. In cluster 2210 (29%) patients showed slightly elevated concentrations of hs‐cTnT and NT‐proBNP 30 days after the index ACS, which subsequently decreased. One year after the index ACS, patients in cluster 2 reached similar concentrations as patients in cluster 1. In cluster 3241 (33%) patients had elevated concentrations of all 4 biomarkers 30 days after the index ACS, which gradually decreased during the months thereafter, but remained elevated compared with clusters 1 and 2. Patients in cluster 3 had temporal biomarker patterns similar to those with a repeat ACS during the first year after the index event. Notably, the clustering algorithm did not identify a cluster of patients characterized by increasing concentrations of hs‐cTnT, NT‐proBNP, or hs‐CRP during follow‐up.
Figure 1. Clusters' biomarker trajectories during the first year following index ACS.
Biomarker trajectories of repeated measurements‐based clusters during the first year following the index ACS. The solid lines depict the average biomarker evolutions in each cluster separately. The dashed lines represent the 95% CI. Samples taken within 30 days after index ACS or repeat ACS are excluded from analyses. ACS indicates acute coronary syndrome; GDF‐15, growth differentiation factor 15; hs‐CRP, high‐sensitivity C‐reactive protein; hs‐cTnT, high‐sensitivity cardiac troponin T; and NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide.
| Cluster 1 | Cluster 2 | Cluster 3 | P value* | Re‐ACS cluster | P value† | |
|---|---|---|---|---|---|---|
| n=272 | n=210 | n=241 | n=31 | |||
| First measurement >30 d after the index ACS‡ | ||||||
| hs‐cTnT, ng/L | 8.0 (6.0–11.0) | 12.0 (9.0–17.0) | 19.0 (15.0–28.0) | <0.001§ | 19.0 (12.0–25.5) | 0.321 |
| NT‐proBNP, pmol/L | 16.0 (9.0–28.0) | 47.0 (27.0–87.5) | 76.0 (42.0–153.0) | <0.001§ | 44.0 (22.5–109.5) | 0.018§ |
| hs‐CRP, mg/L | 1.3 (0.7–2.5) | 1.2 (0.6–2.5) | 2.7 (1.1–5.7) | <0.001§ | 1.3 (0.6–4.4) | 0.056 |
| GDF‐15, pg/mL | 1047.5 (801.5–1325.2) | 896.5 (699.5–1139.0) | 1875.0 (1435.0–2736.0) | <0.001§ | 1383.0 (1061.5–2288.5) | 0.029§ |
| Estimated concentration at 30 d after the index ACS‡ | ||||||
| hs‐cTnT, ng/L | 8.1 (6.0–10.6) | 11.1 (8.9–14.7) | 18.1 (14.1–23.4) | <0.001§ | 19.0 (12.1–22.5) | 0.427 |
| NT‐proBNP, pmol/L | 13.4 (7.7–22.8) | 30.6 (17.7–49.2) | 58.2 (35.2–120.1) | <0.001§ | 38.2 (21.5–99.8) | 0.069 |
| hs‐CRP, mg/L | 1.4 (0.8–2.3) | 1.2 (0.7–2.0) | 2.4 (1.2–4.6) | <0.001§ | 1.8 (1.0–3.3) | 0.106 |
| GDF‐15, pg/mL | 1040 (830–1314) | 901 (718–1158) | 1872 (1443–2573) | <0.001§ | 1536 (1083–2375) | 0.120 |
| Estimated rate of change 30‐d to 1‐y post‐ACSǁ | ||||||
| hs‐cTnT, % change /mo | −1.3 (1.9) | −5.0 (2.3) | −2.5 (2.2) | <0.001§ | −2.2 (1.8) | 0.452 |
| NT‐proBNP, % change/mo | −3.4 (3.3) | −9.8 (3.5) | −6.0 (4.1) | <0.001§ | −4.4 (4.7) | 0.048§ |
| hs‐CRP, % change/mo | −0.4 (2.9) | −0.5 (3.0) | −1.5 (3.1) | <0.001§ | −1.6 (2.7) | 0.781 |
| GDF‐15, % change/mo | 0.6 (1.4) | 0.3 (1.4) | −0.1 (1.4) | <0.001§ | 0.6 (1.7) | 0.010§ |
| Estimated concentration at 1 y after the index ACS‡ | ||||||
| hs‐cTnT, ng/L | 6.8 (5.2–9.4) | 6.6 (4.8–8.7) | 13.3 (10.0–19.2) | <0.001§ | 13.9 (8.9–17.6) | 0.702 |
| NT‐proBNP, pmol/L | 8.8 (5.1–15.5) | 9.8 (5.3–18.1) | 31.9 (16.1–60.0) | <0.001§ | 32.9 (9.8–58.0) | 0.427 |
| hs‐CRP, mg/L | 1.4 (0.7–2.2) | 1.1 (0.7–2.0) | 2.1 (1.0–3.8) | <0.001§ | 1.7 (0.8–2.9) | 0.099 |
| GDF‐15, pg/mL | 1061 (851–1470) | 945 (719–1179) | 1806 (1392–2680) | <0.001§ | 1668 (1149–2574) | 0.383 |
ACS indicates acute coronary syndrome; GDF‐15, growth differentiation factor 15; hs‐CRP, high‐sensitivity C‐reactive protein; hs‐cTnT, high‐sensitivity cardiac troponin T; and NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide.
*
Cluster 1 vs cluster 2 vs cluster 3.
†
Cluster 3 vs re‐ACS cluster.
‡
Biomarker concentrations are presented as median (25th to 75th percentile).
§
P < 0.05 is considered statistically significant.
ǁ
Rate of change in biomarker concentration is presented as mean (SD).
Cluster Membership in Relation to Clinical Characteristics
Table 2 and Figure S3 present an overview and summary of the baseline characteristics of the patients in the separate clusters. The patients in cluster 3 had the poorest clinical profile, with higher age, greater prevalence of cardiovascular risk factors (particularly diabetes and hypertension) and cardiovascular disease, higher GRACE risk scores, and lower estimated glomerular filtration rate than the patients in clusters 1 and 2. Overall, the clinical characteristics of patients in cluster 3 were similar to those of patients with a repeat ACS, although patients in cluster 3 were slightly older, had lower systolic blood pressure, and were more frequently diagnosed with ST‐segment–elevation myocardial infarction. Patients in clusters 1 and 2 had a similar prevalence of cardiovascular risk factors but differed in their cardiovascular history and diagnosis at presentation. Specifically, patients in cluster 1 more often had a history of myocardial infarction and percutaneous coronary intervention and were more likely to present with non–ST‐segment–elevation myocardial infarction or unstable angina pectoris compared with cluster 2.
| Cluster 1 | Cluster 2 | Cluster 3 | P value* | Re‐ACS cluster | P value† | |
|---|---|---|---|---|---|---|
| n=272 | n=210 | n=241 | n=36 | |||
| Age, y, mean (SD) | 59.7 (9.6) | 56.8 (9.0) | 68.5 (9.6) | <0.001‡ | 64.1 (12.4) | 0.013‡ |
| Male sex, n (%) | 214 (78.7) | 172 (81.9) | 183 (75.9) | 0.303 | 27 (75.0) | 1.000 |
| White race, n (% ) | 256 (95.5) | 200 (95.7) | 223 (93.7) | 0.547 | 36 (100.0) | 0.248 |
| Cardiovascular risk factors | ||||||
| Diabetes, n (%) | 51 (18.8) | 28 (13.3) | 86 (35.7) | <0.001‡ | 12 (33.3) | 0.930 |
| Hypertension, n (%) | 147 (54.0) | 99 (47.1) | 157 (65.1) | <0.001‡ | 18 (50.0) | 0.116 |
| Hypercholesterolemia, n (%) | 137 (50.4) | 99 (47.1) | 122 (50.6) | 0.715 | 17 (47.2) | 0.840 |
| Current smoker, n (%) | 114 (41.9) | 104 (49.5) | 74 (30.7) | <0.001‡ | 15 (41.7) | 0.262 |
| Body mass index, kg/m2, mean (SD) | 28.0 (4.5) | 27.0 (3.6) | 28.2 (5.2) | 0.017‡ | 27.2 (3.6) | 0.267 |
| History of cardiovascular disease | ||||||
| Myocardial infarction, n (%) | 85 (31.2) | 32 (15.2) | 77 (32.1) | <0.001‡ | 11 (30.6) | 1.000 |
| CABG, n (%) | 26 (9.6) | 5 (2.4) | 37 (15.4) | <0.001‡ | 9 (25.0) | 0.226 |
| PCI, n (%) | 97 (35.7) | 34 (16.3) | 64 (26.6) | <0.001‡ | 11 (30.6) | 0.762 |
| Stroke, n (%) | 15 (5.5) | 12 (5.7) | 34 (14.1) | 0.001‡ | 8 (22.2) | 0.309 |
| Peripheral arterial disease, n (%) | 15 (5.5) | 7 (3.3) | 32 (13.3) | <0.001‡ | 8 (22.2) | 0.242 |
| Chronic heart failure, n (%) | 4 (1.5) | 2 (1.0) | 8 (3.3) | 0.149 | 2 (5.6) | 0.848 |
| Presentation on admission | ||||||
| GRACE risk score at discharge, mean (SD) | 94.7 (27.2) | 83.7 (23.0) | 112.8 (30.8) | <0.001‡ | 119.9 (42.4) | 0.227 |
| Systolic blood pressure, mm Hg, mean (SD) | 139.3 (27.3) | 137.4 (24.7) | 140.5 (28.0) | 0.479 | 150.8 (22.9) | 0.037‡ |
| eGFR, mL/min per 1.73 m2, mean (SD) | 83.4 (19.7) | 85.3 (19.8) | 72.0 (22.6) | <0.001‡ | 75.1 (17.2) | 0.438 |
| Diagnosis | ||||||
| STEMI, n (%) | 110 (40.4) | 144 (68.6) | 125 (51.9) | <0.001‡ | 11 (30.6) | 0.027‡ |
| NSTEMI, n (%) | 122 (44.9) | 55 (26.2) | 91 (37.8) | <0.001‡ | 20 (55.6) | 0.064 |
| UAP, n (%) | 40 (14.7) | 11 (5.2) | 25 (10.4) | 0.004‡ | 5 (13.9) | 0.730 |
| CAG performed, n (%) | 267 (98.2) | 206 (98.1) | 223 (92.5) | 0.001‡ | 33 (91.7) | 1.000 |
| PCI performed, n (%) | 222 (84.1) | 187 (93.5) | 189 (85.5) | 0.007‡ | 28 (87.5) | 0.977 |
| Discharge medication | ||||||
| Aspirin, n (%) | 267 (98.2) | 207 (98.6) | 229 (95.0) | 0.036‡ | 32 (91.4) | 0.633 |
| P2Y12 inhibitor, n (%) | 265 (97.4) | 204 (97.1) | 227 (94.2) | 0.113 | 33 (91.7) | 0.829 |
| Vitamin K antagonist, n (%) | 9 (3.3) | 8 (3.8) | 23 (9.5) | 0.004‡ | 6 (16.7) | 0.312 |
| Statin, n (%) | 264 (97.1) | 208 (99.0) | 229 (95.0) | 0.045‡ | 32 (91.4) | 0.633 |
| Beta blocker, n (%) | 244 (89.7) | 193 (91.9) | 215 (89.2) | 0.598 | 31 (88.6) | 1.000 |
| ACE inhibitor or ARB, n (%) | 217 (79.8) | 178 (84.8) | 204 (84.6) | 0.235 | 32 (91.4) | 0.419 |
Proportion missing values was <0.3% except for ethnicity (1.1%), body mass index (0.5%), eGFR (2.0%), and PCI performed (5.5%). ACE indicates angiotensin‐converting enzyme; ARB, angiotensin receptor blocker; CABG, coronary artery bypass graft; CAG, coronary angiogram; eGFR, estimate glomerular filtration rate; GRACE, Global Registry of Acute Coronary Events; NSTEMI, non–ST‐segment–elevation myocardial infarction; PCI, percutaneous coronary intervention; STEMI, ST‐segment–elevation myocardial infarction; and UAP, unstable angina pectoris.
*
Cluster 1 vs cluster 2 vs cluster 3.
†
Cluster 3 vs cluster re‐ACS.
‡
P < 0.05 is considered statistically significant.
Cluster Membership in Relation to Long‐Term Survival
Over a median follow‐up of 9.1 (25th to 75th percentile, 7.3–10.4) years after the 1‐year landmark, 141 (18.6%) patients reached the primary end point of all‐cause death. Patients in cluster 3 had a significantly higher cumulative incidence of all‐cause death (cumulative incidence, 39% at 10.4 years) compared with those in cluster 1 (14%) and cluster 2 (9%), who had a similar prognosis (Figure 2). The long‐term cumulative incidence of all‐cause death of patients in cluster 3 was numerically higher than those who experienced a repeat ACS in the first year after the index event (26%). These findings were largely confirmed by the accelerated failure time models (Table 3). After correcting for multiple cardiovascular risk factors, the survival time ratio for cluster 3 was 0.64 (95% CI, 0.44–0.93; P=0.018) relative to cluster 1 and 0.71 (95% CI, 0.39–1.32; P=0.281) relative to the patients with a repeat ACS, indicating a 36% and 29% lower expected survival time, respectively. The adjusted survival times of patients in clusters 1 and 2 were similar, as expressed by an adjusted survival time ratio of 1.04 (95% CI, 0.64–1.67; P=0.882) for cluster 2 relative to cluster 1.
Figure 2. Association between clusters based on repeated biomarker measurements and long‐term all‐cause death.
ACS indicates acute coronary syndrome.
| Model 1: unadjusted | Model 2: adjusted for age | Model 3: adjusted for age and sex | Model 4: adjusted for age, sex, and CVD risk factors† | |||||
|---|---|---|---|---|---|---|---|---|
| STR (95% CI)* | P value | STR (95% CI)* | P value | STR (95% CI)* | P value | STR (95% CI)* | P value | |
| Cluster 1 | Ref | Ref | Ref | Ref | ||||
| Cluster 2 | 1.31 (0.80–2.13) | 0.282 | 1.11 (0.69–1.77) | 0.668 | 1.10 (0.69–1.75) | 0.696 | 1.04 (0.64–1.67) | 0.882 |
| Cluster 3 | 0.34 (0.23–0.50) | <0.001‡ | 0.60 (0.42–0.87) | 0.007‡ | 0.60 (0.42–0.87) | 0.007‡ | 0.64 (0.44–0.93) | 0.018‡ |
| Re‐ACS cluster | 0.50 (0.26–0.95) | 0.034‡ | 0.71 (0.38–1.32) | 0.279 | 0.72 (0.39–1.34) | 0.307 | 0.80 (0.43–1.50) | 0.495 |
| Re‐ACS cluster | Ref | Ref | Ref | Ref | ||||
| Cluster 3 | 0.67 (0.35–1.27) | 0.219 | 0.81 (0.43–1.51) | 0.504 | 0.80 (0.43–1.49) | 0.474 | 0.71 (0.39–1.32) | 0.281 |
ACS indicates acute coronary syndrome; CVD, cardiovascular disease; and STR, survival time ratio.
*
STR produced by accelerated failure time models can be interpreted as the average survival time since the 1‐year landmark of a cluster relative to a reference cluster.
†
Diabetes, hypertension, hypercholesterolemia, smoking status, body mass index, diagnosis of index ACS, history of coronary artery disease, history of stroke, and estimated glomerular filtration rate.
‡
P < 0.05 is considered statistically significant.
Clustering Based on Single 1‐Year Biomarker Estimates
As described above, patients in clusters 1 and 2 showed similar biomarker concentrations 1 year after the index event, which were considerably lower than those of patients in cluster 3. No cluster was identified with increasing biomarker concentrations during 1‐year follow‐up. Therefore, we conducted patient clustering again using individual patients' linear mixed‐effects –based estimates (see Methods section) of biomarker concentrations at 1 year as a single value for each biomarker (Table 1). As expected, this analysis resulted in 2 clusters of 427 and 296 patients with low and high estimated biomarker concentrations at 1 year, respectively. These low and high biomarker clusters largely, but not entirely, overlapped with the combined clusters 1 and 2, and cluster 3, respectively (see Figure 3). The estimated 1‐year biomarker concentrations of patients in the high biomarker cluster were similar to those of patients with a repeat ACS, except for hs‐CRP, which was lower in the patients with repeat ACS (see Figure S4). Patients with high estimated biomarker concentrations had significantly worse death outcomes than patients with low biomarker concentrations. The cumulative incidence at 10.4 years was 37% versus 10% (P<0.001), respectively, and the adjusted survival time ratio for patients in the high biomarker cluster was 0.56 (95% CI, 0.39–0.82; P=0.003) (see Figure S5 and Table S1) relative to patients in the low biomarker cluster.
Figure 3. Reclassification diagram.
The low biomarker cluster based on single 1‐year biomarker estimates coincides mostly with the combined clusters 1 and 2 on the basis of repeated biomarker measurements. The high biomarker cluster based on single 1‐year biomarker coincides mostly with cluster 3 on the basis of repeated biomarker measurements.
DISCUSSION
We conducted a comprehensive cluster analysis of repeatedly measured hs‐cTnT, NT‐proBNP, hs‐CRP, and GDF‐15 in the year following an ACS and studied their associations with clinical characteristics and long‐term mortality risk after the 1‐year landmark. Based on their temporal biomarker patterns, this study identified 3 patients with subphenotypes of low‐risk post‐ACS. These subphenotypes displayed significant differences in their longitudinal biomarker profile, clinical characteristics, and long‐term mortality risk. Patients with persistently elevated biomarker concentrations had the worst prognosis, even compared with the apparent high‐risk patients who experienced a repeat ACS in the first year.17 Furthermore, no subphenotype characterized by increasing biomarker concentrations was identified.
Patients with high concentrations of hs‐cTnT, NT‐proBNP, hs‐CRP, and GDF‐15 at baseline and during follow‐up showed the highest risk of death. This finding is consistent with our previous report on BIOMArCS,5 as well as other studies that have linked elevated concentrations of these biomarkers to an increased risk of adverse cardiovascular outcomes in patients with ACS.3, 5, 25, 26, 27 Patients who experienced a repeat ACS during the first year had similar biomarker patterns as patients in cluster 3. Hence, based on their longitudinal biomarker pattern during the first year after the index event, patients in cluster 3 had an increased risk of a repeat ACS but ultimately remained event free. Nowadays, it is still a considerable challenge to predict the exact timing of the event in high‐risk individuals, despite the increase in knowledge on risk factors, including biomarkers, for the occurrence of an ACS. The uncertainty surrounding the timing of an ACS occurrence is disconcerting but may be inherent to the syndrome. Nevertheless, patients with a repeat ACS had better long‐term prognosis than patients in cluster 3, which could be explained by the fact that patients with a repeat ACS were younger at the time of the index event and may have received intensified management following the repeat event.
Although the temporal biomarker pattern of patients in cluster 2 differed from that of cluster 1, both clusters had similar long‐term prognoses. Specifically, hs‐cTnT and NT‐proBNP concentrations were elevated at baseline in cluster 2 compared with cluster 1 and showed a steeper negative slope in the months thereafter until they reached similar concentrations at 1 year. This prolonged elevation of biomarker concentrations observed in cluster 2 might reflect a greater degree of myocardial damage and dysfunction after the index ACS, which is consistent with the fact that cluster 2 composed mostly of patients with ST‐segment–elevation myocardial infarction.28, 29 Nonetheless, it is important to highlight that these findings also suggest that once the 1‐year landmark is reached, the patient's prognosis primarily hinges on the continued elevation of biomarkers, while the trajectory leading up to that point appears to be of lesser significance. Consequently, our results also indicate that highly frequent blood sampling during the first year of follow‐up may not provide additional incremental prognostic information above a single measurement in clinical practice. Whether this also holds for cardiovascular adverse clinical outcomes requires further research.
In a previous report, we showed that BIOMArCS patients with a repeat ACS were charaterized by systematically elevated biomarker concentrations during follow‐up and not by increasing concentrations.5 The current analysis also did not reveal a cluster of patients with increasing biomarker concentrations. In fact, only 3% of our patients showed a relative hs‐cTnT increase of >25% during the 30‐day to 1‐year study period (based on the linear mixed‐effects model; data not shown). The findings in BIOMArCS contrast with previous reports by White et al,30 Everett et al,31 Cavender et al,32 and Patel et al,33 who reported a meaningful increase in cardiac troponin in up to 23% of patients with coronary artery disease and increasing concentrations were associated with a worse prognosis in all 4 studies. Indeed, increasing cardiac troponin concentrations are disconcerting, as these are potentially related to ongoing myocyte necrosis, chronic processes related to left ventricular hypertrophy,34, 35 increasing renal dysfunction,36 or poor diabetes control.37 We cannot entirely explain the absence of this phenotype in BIOMArCS, and the significant disparity in prevalence cannot be attributed solely to variations in sample size. Possibly, individuals choosing to participate in a highly frequent blood sampling study, such as BIOMArCS, might represent a selected subset of healthier patients with ACS with relatively high treatment compliance that could positively influence biomarker concentrations. On the other hand, we consider the regression to the mean phenomenon a serious limitation of previous studies that investigated change based on 2 measurements only.30, 31, 32, 33
The current study shows that subphenotypes with differences in prognosis could be identified on the basis of repeated measurements of a set of biomarkers previously associated with cardiovascular disease. These findings are indicative of the potential utility of cardiovascular biomarkers in future clinical applications, such as personalized prognostication and patient management. For instance, patients with persistent high cardiovascular biomarker concentrations following ACS may benefit from more intensive monitoring, lower thresholds for diagnostic testing, and earlier intervention. In contrast, patients with a favorable biomarker profile can be reassured and potentially be followed‐up less intensively. While clinical guidelines acknowledge the additional prognostic value of individual and combined biomarkers, their routine use for prognostication and therapeutic decision making has not yet been recommended.1, 38, 39 Ideally, cardiovascular biomarkers would be used as a continuous, longitudinal measure in a dynamic cardiovascular risk prediction tool that incorporates sex, age, and clinical features. Further research is needed to investigate the incremental value of more extensive biomarker assays, such as multimarker assays currently used in heart failure populations,10 and to determine whether these carry the potential to enhance ACS prognostication and personalized risk assessment. Furthermore, since noncardiovascular deaths will most likely not be preventable with cardiovascular secondary prevention strategies, further research is warranted to investigate whether repeated biomarker measurements hold promise in distinguishing patients with differing long‐term prognoses for cardiovascular adverse events.
Our study has several strengths. First, this is the first study that investigated subphenotypes of patients following ACS on the basis of the temporal evolution of multiple cardiovascular biomarkers that were repeatedly measured at high frequency. Most biomarker studies classifying patients following ACS have explored 1 cross‐sectional measurement only, or studied changes using just 2 repeated measurements several months apart.30, 31, 32, 33, 40 Our data in combination with the statistical techniques (ie, based on linear mixed‐effect models) overcomes the regression to the mean phenomenon, a well‐known limitation of observational studies investigating change.41 Second, complete follow‐up for >10 years enabled linking the ACS subphenotypes to long‐term prognosis.
Some limitations also need to be acknowledged. First, the BIOMArCS study comprises a predominantly White population, and caution should be exercised when generalizing our findings to other ethnic groups. Second, during the inclusion period of BIOMArCS, there was a notable increase in the adaptation of more potent P2Y12 inhibitors, namely, ticagrelor and prasugrel, which may affect patients' risk of adverse outcomes. Nonetheless, type of P2Y12 inhibitor was not associated with the risk of a repeat ACS within the first year after the index ACS or long‐term mortality risk during follow‐up in the current study (potent P2Y12 inhibitors versus clopidogrel, P=0.351 and P=0.674, respectively). Third, while all‐cause death is an important clinical outcome, it may not entirely represent the most clinically relevant outcomes from a therapeutic perspective, given that noncardiovascular deaths will most likely not be preventable with cardiovascular secondary prevention strategies. Fourth, clinical and imaging indices of cardiac function, atherosclerotic plaque characteristics, and the degree of coronary artery disease were not routinely measured in this patient population, making it difficult to comment on the combined prognostic value of these indices with cardiovascular biomarkers. Finally, 13 of the 10 606 hs‐cTnT measurements were considered outliers and were excluded from analysis. Nonetheless, subclinical events, such as a subclinical myocardial infarction, cannot be excluded.
In conclusion, the current study demonstrates the ability to identify patients with subphenotypes of post‐ACS based on repeated measurements of established cardiovascular biomarkers. These phenotypes show significant differences in all‐cause death during long‐term follow‐up, with patients exhibiting persistently elevated biomarker concentrations having the worst outcome, regardless of whether they experienced a repeat ACS during the first year. Our findings highlight the valuable role of blood biomarkers, assessed through single or repeated measurements, in long‐term ACS prognostication and personalized risk assessment.
APPENDIX
BIOMArCS Investigators
Erasmus MC, University Medical Center Rotterdam, Rotterdam, the Netherlands: K Martijn Akkerhuis, Marie de Bakker, Eric Boersma, Isabella Kardys, Yolanda B. de Rijke, Ron H. van Schaik, and Niels T.B. Scholte. Ismail Aksoy (Admiraal de Ruyter Hospital, Goes, the Netherlands), Folkert W Asselbergs (Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, The Netherlands and Health Data Research UK and Institute of Health Informatics, University College London, London, United Kingdom), G Etienne Cramer (Radboud University Medical Center Nijmegen, the Netherlands), Pim van der Harst (University Medical Center Utrecht, the Netherlands), Walter Hermans (Elizabeth‐Tweesteden Hospital, Tilburg, the Netherlands), S Hong Kie The (Treant Zorggroep, Emmen, the Netherlands), Bas Kietselaer (Mayo Clinic, Rochester, MN), Boudewijn Krenning (Erasmus MC, University Medical Center Rotterdam, Rotterdam, the Netherlands; and Franciscus Gasthuis & Vlietland, Rotterdam, the Netherlands), Timo Lenderink (Zuyderland Hospital, Heerlen, the Netherlands), Arthur Maas (Gelre Hospital, Zutphen, the Netherlands), Rohit M. Oemrawsingh (Albert Schweitzer Ziekenhuis, Dordrecht, the Netherlands), Anton J. Oude Ophuis (Canisius‐Wilhelmina Hospital, Nijmegen, the Netherlands), Eelko Ronner (Reinier de Graaf Hospital, Delft, the Netherlands), Carl Schotborgh (HagaZiekenhuis, Den Haag, the Netherlands), Victor A. Umans (Noordwest Ziekenhuisgroep, Alkmaar, the Netherlands), Alexander J. Wardeh (Haaglanden Medisch Centrum, Den Haag, the Netherlands), Robbert J. de Winter (Amsterdam University Medical Centers, University of Amsterdam, Amsterdam, The Netherlands).
Sources of Funding
Roche Diagnostics provided assays and kits in support of this work. COBAS C, COBAS E, and ELECSYS are trademarks of Roche. The BIOMArCS study was supported and funded by The Netherlands Heart Foundation, grant number 2007B012 (Utrecht, The Netherlands), The Netherlands Heart Institute–Interuniversity Cardiology Institute of Netherlands, project number 071.01 (Utrecht, The Netherlands), and the Working Group of Cardiovascular Research Netherlands, all of which are noncommercial funding bodies. An unrestricted research grant was further obtained from Eli Lilly, the Netherlands.
Disclosures
M de Bakker is supported by the Jaap Schouten Foundation. Dr Scholte is supported by a grant from the Dutch Research Council (NWO), grant number 628.011.214 (STRAP). Dr Asselbergs is supported by UCL Hospitals NIHR Biomedical Research Centre. The remaining authors have no disclosures to report.
Footnotes
This manuscript was sent to Amgad Mentias, MD, Associate Editor, for review by expert referees, editorial decision, and final disposition.
Supplemental Material is available at Supplemental Material
For Sources of Funding and Disclosures, see page 10.
Supplemental Material
Data S1
Table S1
Figures S1–S5
- Download
- 1002.70 KB
References
1.
Knuuti J, Wijns W, Saraste A, Capodanno D, Barbato E, Funck‐Brentano C, Prescott E, Storey RF, Deaton C, Cuisset T, et al. 2019 ESC Guidelines for the diagnosis and management of chronic coronary syndromes: the Task Force for the diagnosis and management of chronic coronary syndromes of the European Society of Cardiology (ESC). Eur Heart J. 2019;41:407–477.
2.
Lindholm D, Lindbäck J, Armstrong PW, Budaj A, Cannon CP, Granger CB, Hagström E, Held C, Koenig W, Östlund O, et al. Biomarker‐based risk model to predict cardiovascular mortality in patients with stable coronary disease. J Am Coll Cardiol. 2017;70:813–826.
3.
Eggers KM, Lindahl B. Prognostic biomarkers in acute coronary syndromes: risk stratification beyond cardiac troponins. Curr Cardiol Rep. 2017;19:29.
4.
Lyngbæk S, Winkel P, Gøtze JP, Kastrup J, Gluud C, Kolmos HJ, Kjøller E, Jensen GB, Hansen JF, Hildebrandt P, et al. Risk stratification in stable coronary artery disease is possible at cardiac troponin levels below conventional detection and is improved by use of N‐terminal pro‐B‐type natriuretic peptide. Eur J Prev Cardiol. 2014;21:1275–1284.
5.
Oemrawsingh RM, Akkerhuis KM, de Mulder M, Umans VA, Kietselaer B, Schotborgh C, Ronner E, Lenderink T, Liem A, Haitsma D, et al. High‐frequency biomarker measurements of troponin, NT‐proBNP, and C‐reactive protein for prediction of new coronary events after acute coronary syndrome. Circulation. 2019;139:134–136.
6.
Marston NA, Bonaca MP, Jarolim P, Goodrich EL, Bhatt DL, Steg PG, Cohen M, Storey RF, Johanson P, Wiviott SD, et al. Clinical application of high‐sensitivity troponin testing in the atherosclerotic cardiovascular disease framework of the current cholesterol guidelines. JAMA Cardiol. 2020;5:1255–1262.
7.
Pandey A, Patel KV, Vongpatanasin W, Ayers C, Berry JD, Mentz RJ, Blaha MJ, McEvoy JW, Muntner P, Vaduganathan M, et al. Incorporation of biomarkers into risk assessment for allocation of antihypertensive medication according to the 2017 ACC/AHA High Blood Pressure Guideline: a pooled cohort analysis. Circulation. 2019;140:2076–2088.
8.
James SK, Lindbäck J, Tilly J, Siegbahn A, Venge P, Armstrong P, Califf R, Simoons ML, Wallentin L, Lindahl B. Troponin‐T and N‐terminal pro‐B‐type natriuretic peptide predict mortality benefit from coronary revascularization in acute coronary syndromes: a GUSTO‐IV substudy. J Am Coll Cardiol. 2006;48:1146–1154.
9.
Morrow DA, Cannon CP, Rifai N, Frey MJ, Vicari R, Lakkis N, Robertson DH, Hille DA, DeLucca PT, DiBattiste PM, et al. Ability of minor elevations of troponins I and T to predict benefit from an early invasive strategy in patients with unstable angina and non‐ST elevation myocardial infarction: results from a randomized trial. JAMA. 2001;286:2405–2412.
10.
de Lange I, Petersen TB, de Bakker M, Akkerhuis KM, Brugts JJ, Caliskan K, Manintveld OC, Constantinescu AA, Germans T, van Ramshorst J, et al. Heart failure subphenotypes based on repeated biomarker measurements are associated with clinical characteristics and adverse events (bio‐SHiFT study). Int J Cardiol. 2022;364:77–84.
11.
Stienen S, Ferreira JP, Kobayashi M, Preud'homme G, Dobre D, Machu J‐L, Duarte K, Bresso E, Devignes M‐D, López N. Enhanced clinical phenotyping by mechanistic bioprofiling in heart failure with preserved ejection fraction: insights from the MEDIA‐DHF study (The Metabolic Road to Diastolic Heart Failure). Biomarkers. 2020;25:201–211.
12.
Tromp J, Ouwerkerk W, Demissei BG, Anker SD, Cleland JG, Dickstein K, Filippatos G, van der Harst P, Hillege HL, Lang CC. Novel endotypes in heart failure: effects on guideline‐directed medical therapy. Eur Heart J. 2018;39:4269–4276.
13.
Brunton‐O'Sullivan MM, Holley AS, Hally KE, Kristono GA, Harding SA, Larsen PD. A combined biomarker approach for characterising extracellular matrix profiles in acute myocardial infarction. Sci Rep. 2021;11:12705.
14.
Mohebi R, McCarthy CP, Gaggin HK, van Kimmenade RRJ, Januzzi JL. Inflammatory biomarkers and risk of cardiovascular events in patients undergoing coronary angiography. Am Heart J. 2022;252:51–59.
15.
Boersma E, Vroegindewey MM, van den Berg VJ, Asselbergs FW, van der Harst P, Kietselaer B, Lenderink T, Oude Ophuis AJ, Umans V, de Winter RJ, et al. Details on high frequency blood collection, data analysis, available material and patient characteristics in BIOMArCS. Data Brief. 2019;27:104750.
16.
Oemrawsingh RM, Akkerhuis KM, Umans VA, Kietselaer B, Schotborgh C, Ronner E, Lenderink T, Liem A, Haitsma D, van der Harst P, et al. Cohort profile of BIOMArCS: the BIOMarker study to identify the Acute risk of a Coronary Syndrome‐a prospective multicentre biomarker study conducted in The Netherlands. BMJ Open. 2016;6:e012929.
17.
Smolina K, Wright FL, Rayner M, Goldacre MJ. Long‐term survival and recurrence after acute myocardial infarction in England, 2004 to 2010. Circ Cardiovasc Qual Outcomes. 2012;5:532–540.
18.
Charrad M, Ghazzali N, Boiteau V, Niknafs A. NbClust: an R package for determining the relevant number of clusters in a data set. J Stat Softw. 2014;61:1–36.
19.
Package NbClust . Version 3.0.1. The Comprehensive R Archive Network. 2022. Accessed October 18, 2023. https://cran.r‐project.org/web/packages/NbClust/NbClust.pdf
20.
Hennig C. Cluster‐wise assessment of cluster stability. Comput Stat Data Anal. 2007;52:258–271.
21.
Hennig C. Dissolution point and isolation robustness: robustness criteria for general cluster analysis methods. J Multivar Anal. 2008;99:1154–1176.
22.
Hartigan JA, Wong MA. Algorithm AS 136: a k‐means clustering algorithm. J R Stat Soc Series C (Appl Stat). 1979;28:100–108.
23.
Li H, Han D, Hou Y, Chen H, Chen Z. Statistical inference methods for two crossing survival curves: a comparison of methods. PLoS One. 2015;10:e0116774.
24.
van Buuren S, Groothuis‐Oudshoorn K. Mice: multivariate imputation by chained equations in R. J Stat Softw. 2011;45:1–67.
25.
Zhang S, Hao P, Li J, Zhang Q, Yin X, Wang J, Chen Y. Prognostic value of growth differentiation factor‐15 in patients with coronary artery disease: a meta‐analysis and systematic review. Front Cardiovasc Med. 2023;10:1054187.
26.
He LP, Tang XY, Ling WH, Chen WQ, Chen YM. Early C‐reactive protein in the prediction of long‐term outcomes after acute coronary syndromes: a meta‐analysis of longitudinal studies. Heart. 2010;96:339–346.
27.
Ottani F, Galvani M, Nicolini FA, Ferrini D, Pozzati A, Di Pasquale G, Jaffe AS. Elevated cardiac troponin levels predict the risk of adverse outcome in patients with acute coronary syndromes. Am Heart J. 2000;140:917–927.
28.
Sarafoff N, Schuster T, Vochem R, Fichtner S, Martinoff S, Schwaiger M, Schömig A, Ibrahim T. Association of ST‐elevation and non‐ST‐elevation presentation on ECG with transmurality and size of myocardial infarction as assessed by contrast‐enhanced magnetic resonance imaging. J Electrocardiol. 2013;46:100–106.
29.
Hjort M, Eggers KM, Lindhagen L, Baron T, Erlinge D, Jernberg T, Marko‐Varga G, Rezeli M, Spaak J, Lindahl B. Differences in biomarker concentrations and predictions of long‐term outcome in patients with ST‐elevation and non‐ST‐elevation myocardial infarction. Clin Biochem. 2021;98:17–23.
30.
White HD, Tonkin A, Simes J, Stewart R, Mann K, Thompson P, Colquhoun D, West M, Nestel P, Sullivan D, et al. Association of contemporary sensitive troponin I levels at baseline and change at 1 year with long‐term coronary events following myocardial infarction or unstable angina: results from the LIPID Study (Long‐Term Intervention with Pravastatin in Ischaemic Disease). J Am Coll Cardiol. 2014;63:345–354.
31.
Everett BM, Brooks MM, Vlachos HE, Chaitman BR, Frye RL, Bhatt DL; BARI 2D Study Group . Troponin and cardiac events in stable ischemic heart disease and diabetes. N Engl J Med. 2015;373:610–620.
32.
Cavender MA, White WB, Jarolim P, Bakris GL, Cushman WC, Kupfer S, Gao Q, Mehta CR, Zannad F, Cannon CP, et al. Serial measurement of high‐sensitivity troponin I and cardiovascular outcomes in patients with type 2 diabetes mellitus in the EXAMINE trial (Examination of Cardiovascular Outcomes with Alogliptin Versus Standard of Care). Circulation. 2017;135:1911–1921.
33.
Patel SM, Qamar A, Giugliano RP, Jarolim P, Marston NA, Park J‐G, Blazing MA, Cannon CP, Braunwald E, Morrow DA. Association of serial high‐sensitivity cardiac troponin T with subsequent cardiovascular events in patients stabilized after acute coronary syndrome: a secondary analysis from IMPROVE‐IT. JAMA Cardiol. 2022;7:1199–1206.
34.
Lyngbakken MN, Aagaard EN, Kvisvik B, Berge T, Pervez MO, Brynildsen J, Tveit A, Steine K, Røsjø H, Omland T. Cardiac troponin I and T are associated with left ventricular function and structure: data from the Akershus Cardiac Examination 1950 Study. Clin Chem. 2020;66:567–578.
35.
Røsjø H, Andreassen J, Edvardsen T, Omland T. Prognostic usefulness of circulating high‐sensitivity troponin T in aortic stenosis and relation to echocardiographic indexes of cardiac function and anatomy. Am J Cardiol. 2011;108:88–91.
36.
Chesnaye NC, Szummer K, Bárány P, Heimbürger O, Magin H, Almquist T, Uhlin F, Dekker FW, Wanner C, Jager KJ, et al. Association between renal function and troponin T over time in stable chronic kidney disease patients. J Am Heart Assoc. 2019;8:e013091.
37.
Šimić S, Svaguša T, Prkačin I, Bulum T. Relationship between hemoglobin A1c and serum troponin in patients with diabetes and cardiovascular events. J Diabetes Metab Disord. 2019;18:693–704.
38.
Grundy SM, Stone NJ, Bailey AL, Beam C, Birtcher KK, Blumenthal RS, Braun LT, de Ferranti S, Faiella‐Tommasino J, Forman DE, et al. 2018 AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA guideline on the management of blood cholesterol: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines. Circulation. 2019;139:e1082–e1143.
39.
Visseren FLJ, Mach F, Smulders YM, Carballo D, Koskinas KC, Bäck M, Benetos A, Biffi A, Boavida JM, Capodanno D, et al. 2021 ESC Guidelines on cardiovascular disease prevention in clinical practice. Eur Heart J. 2021;42:3227–3337.
40.
Eggers KM, Lagerqvist B, Venge P, Wallentin L, Lindahl B. Persistent cardiac troponin I elevation in stabilized patients after an episode of acute coronary syndrome predicts long‐term mortality. Circulation. 2007;116:1907–1914.
41.
Barnett AG, van der Pols JC, Dobson AJ. Regression to the mean: what it is and how to deal with it. Int J Epidemiol. 2004;34:215–220.
Information & Authors
Information
Published In
Copyright
© 2024 The Authors. Published on behalf of the American Heart Association, Inc., by Wiley. This is an open access article under the terms of the Creative Commons Attribution‐NonCommercial‐NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non‐commercial and no modifications or adaptations are made.
Versions
You are viewing the most recent version of this article.
History
Received: 4 September 2023
Accepted: 22 November 2023
Published online: 12 January 2024
Published in print: 16 January 2024
Keywords
Subjects
Authors
Funding Information
The Netherlands Heart Foundation: 2007B012
The Netherlands Heart Institute–Interuniversity Cardiology Institute of Netherlands
Eli Lilly
Metrics & Citations
Metrics
Citations
Download Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Select your manager software from the list below and click Download.
- ABC as aortic valve, Body Mass Index, COVID, and more: a welcome yet challenging update for cardiovascular practitioners, Minerva Cardiology and Angiology, 72, 2, (2024).https://doi.org/10.23736/S2724-5683.24.06554-2
Loading...
View Options
Login options
Check if you have access through your login credentials or your institution to get full access on this article.
Personal login Institutional LoginPurchase Options
Purchase this article to access the full text.
eLetters(0)
eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.
Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.