In this cohort study, patients with suspected CAD or suspected progress of a known CAD, undergoing exercise or pharmacological cardiac stress testing were prospectively enrolled at the University Heart & Vascular Center Hamburg. The techniques for stress testing included stress echocardiography, MPS, or cardiac magnetic resonance tomography (cMRI). Patients were referred for cardiac stress testing by in-house or primary care physicians due to the occurrence or progression of dyspnea, stable angina pectoris, prior positive imaging, abnormal electrocardiogram findings, or for preoperative risk stratification. All study participants were over the age of 18 and provided written informed consent. Patients with a language barrier or those who had undergone surgery or cardiovascular intervention within the past 4 weeks were excluded from the study. Recruitment was conducted between May 31, 2017 and February 14, 2019. All participants provided written informed consent, and the study protocol was approved by the local Ethics Committee (number: PV4023).

Baseline parameters, including age, sex category, cardiovascular risk factors (hypertension, hyperlipoproteinemia, diabetes, family history of CAD, smoking), and preexisting cardiovascular conditions, were assessed by questionnaire and chart review in all participants. Furthermore, blood samples for biobanking were collected before and 1 hour after the stress test. After centrifugation and aliquotation all blood samples were stored at −80°C for later batched analyses.

The primary endpoint of the study was the occurrence of MACE, defined as a composite of all-cause mortality, incident non-fatal myocardial infarction, myocardial revascularization, admission due to cardiac causes, and non-fatal stroke. Patients were followed from the time of study inclusion for up to 3 years by phone call, medical record, mail, or contact to the general doctor to assess any incident events. All reported hospitalizations and events were verified through discharge letters or medical documentation whenever available. For all patients without available follow-up information, the local death register was checked to record all fatal events. If the cause of death could not be provided, the death certificate was obtained from the local public health department. The median follow-up was 2.87 years (95% CI 2.8 to 2.93) and was completed in November 2021. Complete follow-up for any incident event at 30 days was available for 754 patients, at 1 year for 679 patients, at 2 years for 526 patients, and at 3 years for 238 patients.

Concentrations of baseline hs-cTnI, NT-proBNP, hs-CRP, and GDF-15 were measured in blood samples collected immediately before the start of the stress testing (0h). Additionally, serial hs-cTnI concentrations were measured 1 hour after the stress test (1h). Hs-cTnI was determined using the Abbott Architect high-sensitivity troponin I immunoassay (ARCHITECT i2000SR; Abbott Diagnostics) with a limit of detection of 1.9 (range 0 to 50,000 pg/ml) and a reported 10% coefficient of variation (CV) at a concentration of 5.2 pg/ml. NT-proBNP was determined using an electrochemiluminescence immunoassay by Roche Diagnostics on ELECSYS 2010 and Cobas e411 analyzers (range 5 to 35,000 pg/ml, intra- and inter-assay CV of 2.58 and 1.38 pg/ml, respectively), and hs-CRP measurements were conducted using the CRP Vario immunoassay on an Abbott Architect c8000 (intra- and inter-assay CV of 0.93 and 0.83 ng/L). GDF-15 was determined using a quantitative sandwich kit ELISA assay by R&D (range 23.4 to 1500 pg/ml, intra- and inter-assay CV of 2.8% and 6%, respectively).

Stress echocardiography was conducted on a bicycle ergometer or pharmacologically by intravenous infusion of dobutamine or adenosine according to a standard protocol under continuous 12-lead electrocardiogram monitoring and non-invasive blood pressure measurements, while concomitant imaging was performed. Diagnostic endpoints of stress testing were maximum dose or maximum workload defined by the achievement of an age specific target heart rate. In accordance with the guidelines of the European Association of Echocardiography a positive stress test was defined as ischemic or necrotic response in at least 2 adjacent heart wall segments. ^,

MPS was conducted with ^99m Tc sestamibi as a radioactive tracer. Stress was applied by exercise on a bicycle ergometer or pharmacologically by intravenous infusion of regadenoson according to a standard protocol under continuous 12-lead ECG monitoring and non-invasive blood pressure measurements. Imaging was performed following standard protocols and interpreted with the perfusion score described by Hachamovitch et al. A reversible perfusion defect indicated by a summed difference score >3 was considered a positive test.

CMRI acquisitions were electrocardiographically triggered and performed on a 1.5-Tesla MR scanner (Achieva, Philips Medical Systems, Best, The Netherlands). Regadenoson was used for pharmacological stress, followed by imaging conducted according to a standardized protocol. Perfusion abnormalities were assigned to a coronary artery supply territory using the 17-segment model of the American Heart Association. For stress cardiac MRI, a new stress-induced late gadolinium enhancement was considered a positive test.

Continuous variables in baseline characteristics were expressed as the median with interquartile range (IQR) and were compared between positive/negative stress test by the Mann–Whitney U test. Binary variables in baseline characteristics were expressed as absolute and proportions frequencies and were compared between positive/negative stress test with the chi-square test. Error bars (median and IQR) were plotted for the troponin measurements before and 1 hour after the stress test. Wilcoxon signed rank test was computed to compare hs-cTnI levels at admission and after 1 hour, stratified by stress test result. The diagnostic performance to predict a positive stress test was evaluated by calculation of the area under a receiver-operating-curve (AUC) for baseline measurements of NT-proBNP, hs-CRP, GDF-15, and hs-cTnI, as well as hs-cTnI concentrations after 1h. An optimal cutoff was determined by maximization of the Youden Index (Sensitivity + Specificity − 1). The 95% confidence intervals (CI) were generated using bootstrapping with 2,000 bootstrap replicates. Median follow-up time and event rates were calculated using the reverse Kaplan–Meier estimator. Survival curves were estimated using the Kaplan–Meier estimator stratified by biomarker concentration below and above the median. p Values were calculated via the log-rank test. Cox proportional hazards models were calculated to evaluate the association of biomarker concentrations with incident events. Biomarker values were log-transformed due to skewed distribution. Two models were fitted: Model 1, adjusted for SCORE2 variables (age, sex, hypertension, smoking, diabetes, and hyperlipoproteinemia), history of CAD, history of myocardial infarction, and congestive heart failure; and Model 2, additionally adjusted for positive stress test result. Each biomarker was analyzed individually, and all 4 biomarkers were also included simultaneously in a combined model. Subgroup analyses by stress testing modality (stress echocardiography and myocardial perfusion scintigraphy) were conducted for Model 2 including all biomarkers. Subgroup analysis for cardiac MRI was not feasible due to the low number of cases. The proportional hazards assumption was assessed using Schoenfeld residual plots.

Forest plots for MACE and all-cause mortality endpoints were generated from Cox models comparing (I) single-biomarker and cardiac stress test, (II) combined biomarker, and (III) combined biomarker plus cardiac stress test models, each adjusted for SCORE2 variables (age, sex, hypertension, hyperlipoproteinemia, diabetes, and smoking) as well as history of CAD, prior myocardial infarction, and congestive heart failure. Across all models, high p values and flat smoothed lines indicated that the proportional hazards assumption was fulfilled. The association between SCORE2 variables (age, sex, hypertension, hyperlipoproteinemia, diabetes, and smoking) and incident MACE was assessed using a Cox regression model with a cross-validated C-index and Net reclassification index (NRI). For both the C-index and NRI, incremental model performance was evaluated relative to the SCORE2 baseline model. Models included the SCORE2 baseline model alone, SCORE2 plus pathological stress test, SCORE2 plus each log-transformed biomarker individually, and SCORE2 plus pathological stress test and all biomarkers combined. Seven-fold cross-validation was used to estimate performance measures (C-index and NRI). It was developed to avoid the over-optimism of evaluating a model in the same data. Confidence intervals were computed by bootstrapping 500 times the sevenfold cross-validation. The NRI computation was based on event probabilities at 2 years. The category-based NRI was calculated with cut-offs <1%, 1% to <5%, 5% to <10%, and ≥10%. The impact of adding each logarithmized biomarker into the model was evaluated by determining the difference in the C-indices. Statistical analysis was performed using R version 4.3.1.

A total of 822 patients were recruited, and 57 patients had to be excluded due to missing biomarker information or stress test results (Supplementary Figure 2). The median age of the remaining 765 patients was 70 years (IQR 60–76) and 500 (65.4%) were male ( Table 1 ). The stress test was reported as positive in 234 participants and negative in the other 531 participants. MPS was the most common method for stress testing, accounting for 87.7% (671/765) of participants, while stress echocardiography was performed in 11.1% (85/765) and cMRI in 1.2% (9/765) of the cohort.

The study participants demonstrated a high cardiovascular risk profile as 81.5% (609/747) had hypertension, 59% (431/731) had a history of smoking or are actively smoking, 48.4%(359/742) had hyperlipoproteinemia, and 29.9% (93/311) stated a family history of CAD. Participants with a positive stress test report were significantly more likely to have hypertension (89.9% vs 78%, p < 0.001) and preexisting cardiac conditions, including CAD (61.4% vs 39.6%, p < 0.001) and heart failure (17% vs 4.5%, p < 0.001) as well as previous myocardial infarction (36% vs 16.6%, p > 0.001) and previous coronary revascularization (51.1% vs 29.7%, p < 0.001). The participants who received invasive coronary angiography after stress testing were 12.6% (96/764), mainly when the stress test was positive (31.2%, 73/234).

Baseline hs-cTnI concentrations were higher in participants with a positive stress test compared to those with a negative test (8.1 vs 4.5 pg/ml, p < 0.001) and slightly decreased when remeasured 1 hour after stress testing in both groups (7.6 vs 4.4 pg/ml) ( Table 2 , Figure 1 ). GDF-15 and NT-proBNP concentrations measured before the stress test were also significantly higher in patients with a positive stress test (1,834 vs 1,576.8 pg/ml, 399.5 vs 262.4 pg/ml, p < 0.001) while baseline hs-CRP measurements showed no significant difference between the 2 groups (3.2 vs 2.9 mg/L, p = 0.478).

Table 2

Biomarker concentrations before and after stress test (0h and 1h)

	All (n = 765)	Negative stress test (n = 531)	Positive stress test (n = 234)	p Value
Measurement 0h
Hs-cTnI 0h (pg/ml)	5.3 (2.9, 11.9)	4.5 (2.5, 8.9)	8.1 (4.0, 15.9)	<0.001
NT-proBNP 0h (pg/ml)	296 (110.3, 923.7)	262.4 (96.4, 798.5)	399.5 (166.6, 1204.7)	<0.001
Hs-CRP 0h (mg/L)	3.0 (1.2, 7.7)	2.9 (1.2, 7.6)	3.2 (1.3, 7.7)	0.47
GDF-15 0h (pg/ml)	1671.0 (1059.2, 2877.3	1576.8 (964.1, 2828.6)	1834.0 (1184.6, 3076.1)	0.011
Measurement 1h
Hs-cTnI 1h (pg/ml)	5.5 (2.8, 11.1)	4.4 (2.5, 9.2)	7.6 (3.9, 14.8)	<0.001
Hs-cTnI delta 1h (pg/ml)	−0.1 (−0.8, 0.4)	−0.1 (−0.7, 0.4)	−0.2 (−1.2, 0.3)	0.018

Only gold members can continue reading. Log In or Register to continue

Share this:

• X
• Facebook

Related

Related posts:

1. Sex-Based Differences in Cardiovascular Diagnoses, Mortality, and Resource Utilization in Long COVID Hospitalizations: A National Inpatient Sample Analysis
2. Screening of Atrial Fibrillation/Arrhythmia Events in Patients With Abnormal ECHOcardiographic Parameters: The Randomized, Prospective SAFE-ECHO Study Design and Rationale
3. Argatroban for the Treatment of Acute Ischemic Stroke: A Meta-Analysis of Randomized Controlled Trials
4. Impact of Prosthesis–Patient Mismatch After Surgical Mitral Valve Replacement: Systematic Review and Meta‐Analysis

Stay updated, free articles. Join our Telegram channel

Join