Abstract
Background
Chest pain is a common and challenging complaint in emergency departments (EDs), necessitating accurate risk stratification to identify patients at risk for major adverse cardiac events (MACE) while avoiding unnecessary admissions. Several scoring systems have been developed for this purpose, yet their external validity in Middle Eastern populations remains understudied.
Objective
To compare and validate the prognostic accuracy of HEART, TIMI, GRACE, EDACS-ADP, and HET scoring systems in predicting 6-week MACE among patients with chest pain presenting to two tertiary care centers in Isfahan, Iran.
Methods
This retrospective cohort study included adult patients (aged > 18 years) who presented with non-traumatic chest pain to two tertiary referral centers in Isfahan between February and June 2024. Patients’ clinical data, laboratory results, and electrocardiograms (ECGs) were retrieved to calculate standardized cardiac risk scores. The primary outcome was the occurrence of major adverse cardiac events (MACE) within 6 weeks following emergency department (ED) presentation. A 6-week evaluation window was selected based on institutional follow-up protocols, data availability, and existing literature that supports this timeframe as a critical period for early cardiac risk stratification. Diagnostic performance of the risk scores was evaluated using receiver operating characteristic (ROC) curve analysis, including calculation of sensitivity, specificity, positive and negative predictive values, and likelihood ratios at clinically relevant cut-off thresholds.
Results
A total of 274 patients were finally included. Among them 68 (24.8%) met the MACE at presentation or within 6 weeks. The HEART score demonstrated the highest AUC: 0.925 and sensitivity: 97.1%; NPV: 98.18% at cut-off ≤ 3, followed closely by the HET score with AUC: 0.906 and sensitivity: 92.6%; NPV: 95.58% at cut-off ≤ 1. TIMI also performed well in identifying very low-risk patients (AUC: 0.868; sensitivity: 98.5%, NPV: 98.17%, though with limited specificity (26.7%). GRACE and EDACS-ADP showed moderate predictive ability, with AUCs of 0.815 and 0.803, respectively. Performance variations were attributed to differences in population demographics, and study design.
Conclusion
The HEART and TIMI scores at the cut-offs of 3 and 1, respectively demonstrated superior discriminative ability in predicting 6-week MACE in this tertiary care cohort, supporting their use in ED settings for early discharge decisions. HET score also showed utility for ruling out MACE in high-risk patients, however, needs further validation due to its novelty and discrepancies observed among studies. These findings support the local implementation of HEART or TIMI in ED protocols, with further multicenter prospective validation recommended.
Clinical trial number
Not applicable.
Supplementary Information
The online version contains supplementary material available at 10.1186/s12873-025-01327-4.
Keywords: Chest pain, Emergency department, Major adverse cardiac events, HEART, TIMI, GRACE, EDACS-ADP, HET, Risk stratification
Introduction
Chest pain is one of the most common presenting complaints in the emergency department (ED), accounting for 9–10% of annual hospital visits [1, 2]. Its etiologies range from benign causes, such as musculoskeletal pain, to life-threatening conditions, notably acute coronary syndrome (ACS) [3]. ACS represents a spectrum of ischemic heart diseases caused by an abrupt reduction in coronary blood flow, including ST-segment elevation myocardial infarction (STEMI), non-ST-segment elevation myocardial infarction (NSTEMI), and unstable angina (UA). It typically manifests with symptoms such as chest discomfort, pressure, burning sensations, radiating pain, dyspnea, and unexplained fatigue [4].
According to the World Health Organization, coronary artery disease (CAD) accounted for 8.9 million deaths worldwide in 2019, representing 16% of all-cause mortality [5, 6]. Given the potentially fatal consequences of missed ACS diagnoses, accurate and timely triage in the ED is essential [7]. Misdiagnosis may result in adverse outcomes such as myocardial infarction or sudden cardiac death, while overly conservative management can lead to unnecessary admissions and testing, straining healthcare resources and contributing to overdiagnosis [8, 9].
Although only approximately 15% of emergency department (ED) patients presenting with chest pain are ultimately diagnosed with acute coronary syndrome (ACS), the remaining 85% typically have non-cardiac, non-life-threatening conditions [10]. This underscores the need for efficient and reliable diagnostic strategies to guide clinical decision-making in the ED. To address this challenge, several risk stratification models have been developed to facilitate rapid evaluation and disposition of patients with chest pain. These tools assist in categorizing patients into low- and high-risk groups, thereby informing decisions regarding discharge, observation, or further investigation.
Among the most widely studied models are the HEART score (History, Electrocardiogram, Age, Risk factors, and Troponin), the TIMI score (Thrombolysis in Myocardial Infarction), the GRACE score (Global Registry of Acute Coronary Events), the EDACS-ADP (Emergency Department Assessment of Chest Pain – Accelerated Diagnostic Protocol), and the HET score (History, Electrocardiogram, and Troponin). Each scoring system incorporates a combination of weighted clinical variables—including patient demographics, medical history, physical examination findings, ECG changes, and cardiac biomarker levels—to estimate the likelihood of major adverse cardiac events (MACE) [11–17].
These tools are especially useful for identifying low-risk patients, including those with atypical ACS presentations such as NSTEMI or UA, which often lack definitive ECG changes. This facilitates earlier and safer discharge decisions [18]. Notably, HEART, EDACS-ADP, and HET were specifically designed for use in ED, while TIMI and GRACE were initially developed for risk stratification following confirmed ACS but have since been adapted for ED assessment. Several validation studies have demonstrated that these tools—particularly HEART and TIMI—exhibit high sensitivity and negative predictive value (NPV) in predicting MACE [12, 19, 20].
Despite reported negative predictive values (NPVs) frequently exceeding 95%, concerns remain about missed diagnoses and post-discharge complications. This is particularly critical in emergency department (ED) settings, where the acceptable diagnostic error rate is typically less than 1% [21–23]. These concerns are substantiated by recent data indicating that 13.2% of patients presenting with chest pain who revisited the ED unscheduled were subsequently diagnosed with an acute cardiovascular emergency [24].
Additionally, the diagnostic performance of existing risk stratification tools has demonstrated considerable variability across different patient populations and healthcare environments. Notably, none of these tools have been externally validated in an Iranian clinical context.
In response to this gap, the present study aims to compare the predictive performance of five established risk stratification instruments within an Iranian ED population. The broader objective is to assess their external validity, identify the tool most suitable for local implementation, and define optimal threshold values that support safe and timely discharge decisions. We hypothesize that the HEART score will exhibit superior diagnostic accuracy in predicting 6-week major adverse cardiac events (MACE), particularly within a high-acuity, referral-based emergency care setting.
Methods and materials
Study design
This retrospective cohort study was conducted in 2024 and included adult patients presenting with chest pain to the emergency departments of two specialized cardiac referral hospitals in Isfahan, Iran—Shahid Chamran and Khorshid. The primary aim was to evaluate the diagnostic accuracy of several established risk stratification tools in a real-world emergency care context.
Patient exposure and outcomes were assessed over a defined period, allowing for follow-up of clinical endpoints. A post hoc sample size estimation was performed using the observed major adverse cardiac event (MACE) rate and area under the curve (AUC) metrics. The calculated sample size provided adequate statistical power to detect a minimum clinically relevant difference of 0.05 in diagnostic performance between scoring systems.
Sample size
The sample size for this study was pre-calculated based on a cross-sectional study design, aiming for 95% confidence and 80% power. The estimated sample size was calculated to be 282 patients, based on the following assumptions: a confidence interval of 95% (z = 1.94), a power of 80% (β = 0.84), and an estimated event rate for major adverse cardiac events (MACE) of 25%. The margin of error was set to 0.05. However, due to logistical limitations during the study, 274 patients were included in the final analysis.
Study population
This retrospective study reviewed patient records to identify individuals aged 18 years or older who presented to the emergency department (ED) with a primary complaint of chest pain between February 1 and June 30, 2024. All patients were eligible for inclusion regardless of triage level. Patients were excluded if their presentation was primarily due to trauma, if they had a documented hematologic disorder, declined participation, or were unreachable during follow-up communications.
To avoid confounding variables that may influence biomarker interpretation and cardiovascular risk stratification, patients with hematologic disorders were excluded. These conditions, including hematologic malignancies (e.g., leukemia and lymphoma), chronic anemia requiring transfusion, severe thrombocytopenia (platelet count < 50,000/µL), inherited coagulopathies such as hemophilia, sickle cell disease, and myeloproliferative neoplasms, can independently modify inflammatory profiles, troponin kinetics, or hemodynamic responses. Such interference has the potential to mimic acute coronary syndromes or compromise the accuracy of biomarker-driven scoring systems, thus impacting the validity of predictive analyses [25].
Study protocol
Clinical data were retrospectively extracted from the medical records of Shahid Chamran and Khorshid Hospitals. The extracted variables included chest pain characteristics, cardiovascular risk factors, medical history, electrocardiogram (ECG) findings, and laboratory results such as cardiac troponin T (cTnT) and serum creatinine levels. All data were collected using a standardized checklist and entered into a dedicated study database.
Initial clinical data were obtained by a final-year medical intern under supervision. ECG findings relevant to risk score calculation were independently interpreted by two emergency medicine specialists and one cardiologist, all blinded to patient outcomes. Chest pain characteristics were classified by the supervised intern as highly suggestive of acute coronary syndrome (ACS), moderately suggestive, or not suggestive. These classifications corresponded to scores assigned by the HEART and HET systems (2, 1, or 0 points) and were recorded as either positive or negative history in the TIMI score. For EDACS, specific pain-related characteristics were extracted in accordance with its validated scoring criteria.
Cardiovascular risk factors, including smoking history, diabetes mellitus, hypertension, hypercholesterolemia, family history of ischemic heart disease (defined as onset before age 55 in men and 65 in women), the presence of an implantable cardioverter defibrillator (ICD), and previously documented coronary artery disease (CAD), were retrieved from patient records. Supplementary information was obtained through follow-up telephone interviews when required.
Laboratory data, including serum troponin and creatinine levels, were extracted from hospital information systems. High-sensitivity troponin assays were used at Khorshid Hospital, whereas ultra-sensitive kits were employed at Shahid Chamran Hospital. Troponin values were interpreted based on predefined thresholds: values below 0.03 ng/mL or 19 were considered negative; values between 0.03 and 0.08 ng/mL or 19–100 were classified as borderline, contributing 1 point to the HEART and HET scores; and values above 0.08 ng/mL or 100 were interpreted as abnormal. Due to incomplete creatinine testing in the emergency department, values from within the preceding three months were utilized when available, permitting GRACE score calculation only for those.
Electrocardiograms (ECGs) were independently interpreted by two emergency medicine specialists and one cardiologist, all of whom were blinded to outcome data. The interpretations followed standardized criteria. Ischemic changes were defined as new or presumed new ST-segment elevation at the J-point (≥ 0.2 mV in leads V1–V3 or ≥ 0.1 mV in other leads), or ST-segment depression ≥ 0.5 mm in two or more contiguous leads, including reciprocal changes. These findings were scored as 2 points in the HEART and HET scoring systems and classified as ischemic in other risk tools. Non-specific changes, including T-wave inversions ≥ 1 mm, pathological Q-waves exceeding 30 ms, or new left/right bundle branch blocks, were awarded 1 point in the HEART score and recorded as abnormal in other scoring systems. In cases where prior ECGs were unavailable, any detected abnormalities were presumed to be new.
All risk scores were calculated by the researcher using predefined scoring tables (see Appendix 1), with the researcher blinded to both predictive variables and patient outcomes during the six-week follow-up period.
Bias management
To minimize potential sources of bias, all eligible patients presenting with chest pain during the study period were consecutively enrolled on randomly selected days, regardless of triage level or initial clinical suspicion of acute coronary syndrome (ACS). Data extraction was conducted using a standardized checklist to ensure methodological consistency across sites. Risk score calculations were performed by a trained researcher who was blinded to both patient outcomes and other predictive variables. Electrocardiogram (ECG) interpretations were conducted independently by emergency medicine specialists and a cardiologist using predefined criteria, thereby reducing inter-observer variability. Scoring systems were applied uniformly at both participating hospitals to preserve consistency in risk assessment procedures.
Outcome measures
The primary outcome was the occurrence of major adverse cardiac events (MACE), defined as ST-elevation myocardial infarction (STEMI), non-ST-elevation myocardial infarction (NSTEMI), emergency revascularization procedures, cardiovascular mortality, and cardiac arrest. Events were recorded during the index hospital visit or within a six-week follow-up period, selected to align with existing emergency department–based studies evaluating early cardiac risk. Although longer durations may capture delayed presentations, this timeframe was chosen to prioritize clinically actionable decision-making and has been acknowledged as a methodological limitation.
Patients discharged from the emergency department were contacted via telephone to identify any subsequent hospital admissions or cardiac investigations within the designated follow-up period. MACE events were adjudicated by two independent reviewers blinded to scoring outcomes. In the event of disagreement, a third senior clinician facilitated consensus. Adjudication was guided by predefined criteria consistent with international standards.
While established scoring systems apply varied endpoints (e.g., HEART: 3 months; TIMI: 14 days), a 6-week interval was selected based on local practice norms and availability of retrospective data. This timeframe is widely utilized in real-world emergency settings and adequately captures both early and intermediate cardiac events while minimizing recall bias during post-discharge surveillance [26–28].
Ethical considerations
This study received ethical approval from the Ethics Committee of Isfahan University of Medical Sciences (IR.MUI.MED.REC.1403.280). In accordance with local guidelines for retrospective research, informed consent was waived. All patient data were anonymized prior to analysis to ensure confidentiality, and ethical standards adhered to both institutional and national regulations. The investigators declared no conflicts of interest.
Statistical analyses
Statistical analyses were conducted using IBM SPSS Statistics for Windows, Version 26.0 (IBM Corp., Armonk, NY). The distribution of continuous variables was assessed using the Shapiro–Wilk test. Non-normally distributed and ordinal variables were summarized using medians and interquartile ranges (IQR), while normally distributed continuous variables were expressed as means with standard deviations. Group comparisons for ordinal and non-normally distributed variables were performed using the Mann–Whitney U test; categorical variables were compared using the chi-square test, and independent samples t-tests were used for normally distributed continuous variables. Statistical significance was defined as a two-tailed p-value < 0.005.
The diagnostic performance of each risk stratification tool (HEART, TIMI, and GRACE scores) in predicting major adverse cardiac events (MACE) was evaluated using receiver operating characteristic (ROC) curve analysis, and the area under the curve (AUC) was reported. Pairwise comparisons of AUCs were performed using DeLong’s test to assess significant differences in predictive accuracy between the scoring systems. For each model, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and corresponding positive and negative likelihood ratios were calculated at relevant threshold values.
Results
Patient enrollment and primary outcome
A total of 274 patients who presented to the emergency department with chest pain were included in the final analysis. Among them, 68 patients (24.8%) experienced a major adverse cardiac event (MACE) within six weeks. Most of these events were acute myocardial infarctions (AMI, n = 66). 56 patients underwent percutaneous coronary intervention (PCI), 5 required coronary artery bypass grafting (CABG), and 4 deaths were recorded during the follow-up period. Due to missing serum creatinine data, the GRACE score could only be calculated for 163 patients. Among these, 62 patients (38.03%) experienced MACE and 101 did not (Fig. 1).
Fig. 1.
Patient flow chart and breakdown of primary outcome of 30-day major adverse cardiac events (MACE). (AMI: acute myocardial infarction, PCI: percutaneous coronary intervention, CABG: coronary artery bypass grafting)
Baseline characteristics
The mean age of the study population was 53.3 years, and 62% were male. Baseline characteristics and cardiovascular risk factors are presented in Table 1. Patients who experienced MACE within 6-week were more likely to be male and had higher ages, rates of smoking, aspirin use, hypercholesterolemia, and a documented history of coronary artery disease (CAD). These variables showed statistically significant associations with MACE (p < 0.005). In contrast, hypertension, diabetes, family history of ischemic heart disease, and body mass index (BMI < 30) did not differ significantly between groups.
Table 1.
Baseline characteristics and cardiac risk factors of patients with and without 6-week MACE
| Variablesa | Total population (n = 274) | 6-week MACE (n = 68) | No MACE at 6-weeks (n = 206) | P-value |
|---|---|---|---|---|
| Age, mean (SD) | 53.33 | 60.60 | 50.93 | > 0.001 |
| Male gender, n (%) | 170 (62.04%) | 58 (85.29%) | 112 (54.36%) | > 0.001 |
| Smoker, n (%) | 100 (36.49%) | 37 (54.41%) | 63 (30.58%) | 0.001 |
| Hypertension, n (%) | 126 (45.98%) | 36 (52.94%) | 90 (43.68%) | 0.184 |
| Hypercholesterolemia, n (%) | 93 (33.94%) | 34 (50%) | 59 (28.64%) | 0.001 |
| Diabetes, n (%) | 57 (20.8%) | 17 (25%) | 40 (19.41%) | 0.325 |
| Family history, n (%) | 65 (23.72%) | 16 (23.52%) | 49 (23.78%) | 0.846 |
| BMI < 30 | 61 (22.26%) | 18 (26.48) | 43 (20.87%) | 0.336 |
| ASA useb, n (%) | 95 (34.67%) | 34 (50%) | 61 (29.61%) | 0.002 |
| History of CAD (documented coronary stenosis), n (%) | 65 (23.72%) | 29 (42.64%) | 36 (17.47%) | > 0.001 |
a No missing data was observed for the variables listed. All data was complete and available for analysis
b Use of Aspirin Whitin one week before presentation
Risk scores and predictive performance
All five risk stratification tools demonstrated significantly higher median and mean scores in patients who experienced a MACE compared to those who did not. under the adjusted significance threshold (p < 0.005), all stratification tool’s scores were statistically significant (Table 2).
Table 2.
The median and mean scores of patients with and without 6-week MACE by each stratification tools
| Stratification tool | Total population (n = 274) | 6-week MACE (n = 68) | No 6-week MACE (n = 206) | P-value |
|---|---|---|---|---|
| HEART score (0–10) | Mean: 3.40 | Mean: 6.26 | Mean: 2.46 | 0.000 |
| Median: 3 | Median: 7 | Median: 2 | ||
| IQR: (2–4) | IQR: (5–8) | IQR: (1–3) | ||
| HET score (0–6) | Mean: 1.32 | Mean: 3.54 | Mean: 0.59 | 0.000 |
| Median: 1 | Median: 4 | Median: 0 | ||
| IQR: (0–2) | IQR: (2–5) | IQR: (0–1) | ||
| TIMI score (0–7) | Mean: 1.99 | Mean: 3.79 | Mean: 1.40 | 0.000 |
| Median: 2 | Median: 4 | Median: 1 | ||
| IQR: (1–3) | IQR: (3–5) | IQR: (0–2) | ||
| EDACS-ADP b score (0–1) | Mean: 0.46 | Mean: 0.91 | Mean: 0.31 | 0.000 |
| Median: 0 | Median: 1 | Median: 0 | ||
| IQR: (0–1) | IQR: (1–1) | IQR: (0–1) | ||
| GRACE a score (1-384) | Mean: 107.37* | Mean: 131.48* | Mean: 92.56* | 0.000 |
| Median: 104 | Median: 127 | Median: 87 | ||
| IQR: (81–128) | IQR: (105–153) | IQR: (74–111) |
a GRACE score was calculated for 163 patients with available creatinine levels
b EDACS-ADP scores were binarized for analysis: 1 = High Risk, 0 = Low Risk. 91.2% of patients in the MACE group and 31% of those without MACE were classified as high-risk by the EDACS-ADP
Diagnostic accuracy (ROC analysis)
ROC curve analysis confirmed the strong discriminatory ability of all scoring systems in predicting six-week MACE (Fig. 2). As shown in Table 3, the HEART score achieved the highest area under the curve (AUC = 0.925), followed by HET (AUC = 0.906) and TIMI (AUC = 0.868). All AUCs were statistically significant (p < 0.001), reinforcing the diagnostic reliability of these scoring systems.
Fig. 2.
Receiver operating characteristic (ROC) curve of HEART, HET, TIMI, GRACE, EDACS-ADP in predicting 6-week MACE
Table 3.
Area under the curve (AUC) of HEART, HET, TIMI, GRACE, EDACS-ADP
| Stratification tool | AUC (95% CI) |
|---|---|
| HEART | 0.925 (0.885–0.965) |
| HET | 0.906 (0.855–0.957) |
| TIMI | 0.868 (0.820–0.916) |
| GRACE | 0.815 (0.750–0.880) |
| EDACS-ADP | 0.803 (0.747–0.859) |
Pairwise comparison of ROC curves
Pairwise ROC analysis was conducted using the DeLong test to evaluate the discriminatory performance of each scoring system in predicting six-week major adverse cardiac events (MACE). As presented in Table 4, the HEART score exhibited the highest area under the curve (AUC = 0.925), significantly outperforming TIMI (AUC = 0.868; p = 0.0001), GRACE (AUC = 0.815; p < 0.001), and EDACS-ADP (AUC = 0.803; p < 0.0001). These results reinforce HEART’s superior predictive ability for identifying patients at risk of MACE.
Table 4.
Pairwise comparison of ROC curves of HEART, HET, TIMI, GRACE, EDACS-ADP
| Stratification tool Comparison | Differences between areas | P-value |
|---|---|---|
| HEART ~ HET | 0.0194 | 0.2326 |
| HEART ~ TIMI | 0.0568 | 0.0001 |
| HEART ~ EDACS-ADP | 0.122 | < 0.0001 |
| HEART ~ GRACE | 0.110 | < 0.0001 |
| HET ~ TIMI | 0.0374 | 0.0707 |
| HET ~ EDACS-ADP | 0.103 | 0.0007 |
| TIMI ~ EDACS-ADP | 0.0654 | 0.0215 |
| TIMI ~ GRACE | 0.0534 | 0.0242 |
* The Delong test was used for pairwise comparison of the tools, however, due to the missing data the results for the GRACE score should be interpreted with caution
The HET score (AUC = 0.906) also demonstrated high discriminative performance, with no statistically significant difference compared to HEART (p = 0.2326), indicating comparable diagnostic strength. Additionally, HET significantly outperformed EDACS-ADP (p = 0.0007), while its advantage over TIMI was borderline (p = 0.0707). TIMI, in turn, showed significantly greater discriminatory capacity than EDACS-ADP (p = 0.0215). Full comparative statistics are presented in Table 4 and visualized in Fig. 2.
Sensitivity, specificity, predictive values, and likelihood ratios
Two key scoring systems, HEART ≤ 3 and TIMI ≤ 1, demonstrated high negative predictive values (> 98%), making them effective tools for ruling out MACE in emergency department settings. However, no single cutoff point achieved both high sensitivity and specificity, indicating the need for careful risk stratification. While HEART ≤ 3 and TIMI ≤ 1 are reliable for identifying low-risk patients, higher thresholds, such as HET = 2, offer stronger positive likelihood ratios, suggesting increased probability of adverse cardiac events when the score exceeds the threshold (Table 5).
Table 5.
Sensitivity, specificity, negative predictive value (NPV), positive predictive value (PPV), negative likelihood ratio (NLR), positive likelihood ratio (PLR) of key cut-offs for each stratification tool
| Stratification tool | Cut-off | Sensitivity (%) | Specificity (%) | NPV (%) | PPV (%) | PLR | NLR |
|---|---|---|---|---|---|---|---|
| HEART | ≤ 3 | 97.1 | 51.9 | 98.18 | 39.98 | 2.01 | 0.05 |
| ≤ 4 | 86.8 | 78.2 | 94.72 | 56.79 | 3.98 | 0.16 | |
| HET | ≤ 1 | 92.6 | 52.9 | 95.58 | 39.35 | 1.96 | 0.13 |
| ≤ 2 | 80.9 | 91.7 | 93.56 | 76.28 | 9.74 | 0.20 | |
| TIMI | ≤ 1 | 98.5 | 26.7 | 98.17 | 30.72 | 1.34 | 0.05 |
| ≤ 2 | 92.6 | 63.1 | 96.27 | 45.30 | 2.50 | 0.11 | |
| GRACE | ≤ 91 | 91.9 | 55.4 | 95.3 | 40.4 | 2.060 | 0.14 |
| EDACS-ADP | = low Risk | 91.2 | 69.4 | 95.98 | 49.59 | 2.98 | 0.12 |
Net reclassification analysis
To quantify the clinical impact of improved patient stratification, the Net Reclassification Index (NRI) was calculated to compare the performance of the HEART score against the TIMI and GRACE models. Compared to TIMI, the HEART score correctly reclassified − 1.4% more events and reduced incorrect reclassification of non-events by 25.2%, yielding a statistically significant net improvement of 23.7% (p = 0.02).
Discussion
This study evaluated the prognostic accuracy of five widely used chest pain risk stratification tools (HEART, TIMI, GRACE, HET, and EDACS-ADP) in predicting 6-week MACE among patients presenting with chest pain in two Iranian tertiary care hospitals. Table 3 presents the AUCs, which summarize each tool’s overall discriminatory performance by balancing sensitivity and specificity. Based on AUC as an indicator of overall performance, our findings suggest that the HEART, HET, and TIMI scores demonstrated superior discriminatory power, while EDACS and GRACE showed moderate predictive value. However, as each tool’s cut-off value directly impacts sensitivity and NPV, optimal thresholds should be selected based on clinical priorities and individual tool characteristics.
Importantly, the AUC values observed in our study were consistently higher than those reported in previous literature [17, 29, 30]. This discrepancy warrants careful analysis, as it may reflect a true increase in model performance due to the unique characteristics of our population rather than methodological bias. These differences could be attributed to the broader, more heterogeneous nature of our study population—reflective of real-world tertiary referral center settings—which included both low-risk self-referred patients and high-risk physician-referred and EMS-transported cases (e.g., definite STEMI). This may have amplified the scoring systems’ discriminative ability, which should be analyzed in the context of each scoring system.
HEART score
Among all evaluated tools, the HEART score achieved the highest diagnostic performance, with an AUC of 0.925 (95% CI: 0.885–0.965), showing excellent discrimination in ruling out short-term cardiac events. At a cut-off of ≤ 3, the HEART score yielded a sensitivity of 97.1%, specificity of 51.9%, and a NPV of 98.18%, effectively identifying patients at low risk of MACE. Lowering the cut-off to ≤ 2 marginally increased sensitivity but did not improve NPV (remained at 98%), suggesting limited clinical benefit and potentially increasing unnecessary admissions.
Our findings align with existing literature but reveal some differences in performance metrics. Fernando et al. (2019) in a systematic review reported a pooled sensitivity of 95.9% (95% CI: 93.3–97.5%) and specificity of 44.6% (95% CI: 38.8–50.5%) for a HEART score ≤ 4 [29]. While this threshold is one point higher than ours, the comparable sensitivity supports our cut-off selection in prioritizing safety for early discharge. Van den Berg and Body (2018) similarly found a pooled sensitivity of 96.7% and NPV of 97.4–100% for a cut-off of ≤ 3, along with a pooled AUC of 0.81 (95% CI: 0.77–0.84), emphasizing its utility in in reducing unnecessary admissions without compromising safety, especially in resource-constrained EDs [30]. Our slightly higher performance metrics may reflect our elevated baseline MACE rate (24.8%) compared to the 15–17% in international studies.
The choice of cut-off is thus a balance between optimizing early discharge and ensuring patient safety. In our cohort, the cut-off of ≤ 3 offered a reasonable compromise, preserving high sensitivity while improving specificity compared to lower thresholds.
HET score
The HET score, although relatively novel and less widely validated, performed comparably to HEART with an AUC of 0.906 (95% CI: 0.855–0.957), a sensitivity of 92.6%, and an NPV of 95.58% for predicting 6-week MACE at a cut-off of ≤ 1. Raising the threshold (e.g., to 2) significantly increased specificity but reduced sensitivity to 82.4%, limiting its safety for ED use.
Compared to those reported by Löfmark et al. (2023), who first proposed the HET score as a simplified alternative to the HEART score, our results showed better overall performance in terms of AUC. In their study, the HET score demonstrated an AUC of 0.887, slightly higher than the HEART score’s AUC of 0.853—both lower than observed in our study. However, surprisingly, in their study, at a cut-off of < 2, the HET score achieved a sensitivity of 98.7% and an NPV of 99.4% for MACE [17], which is considerably higher sensitivity than what have been observed at same cut-off in our cohort. This discrepancy highlights that high AUC values may not always reflect clinical safety, especially if sensitivity is compromised at standard cut-offs.
The notably higher AUCs observed in this study may reflect the unique characteristics of the cohort, which included patients from heart-referring centers and emergency departments. This design introduced a heterogeneous population comprising both high-risk individuals (frequently transferred via EMS with suspected ST-elevation myocardial infarction [STEMI] or acute coronary syndrome [ACS]) and low-risk patients who self-presented with nonspecific chest pain. Such bimodal risk distribution likely amplified the discriminatory power of the scoring systems, consistent with the spectrum effect—where wide variation in disease likelihood enhances diagnostic accuracy.
Moreover, the exceptional AUCs observed in the HEART and HET scores may be partially attributed to their incorporation of high-yield components such as ECG and troponin values. These parameters are often abnormal in referred STEMI cases, elevating the scores’ sensitivity and reinforcing their effectiveness in stratifying high-risk patients.
Despite the HET score’s promising performance and operational simplicity, its limited validation warrants caution. A threshold of ≤ 1 may be suitable for ruling out MACE in emergency settings, but further international studies are needed before it can be endorsed as a standalone triage tool. Compared to well-established instruments like HEART and TIMI, HET remains in an early phase of validation, and clinicians should be prudent in its clinical application until broader evidence emerges.
TIMI score
Originally developed for prognostication in patients with confirmed ACS, the TIMI score demonstrated robust performance in our cohort for identifying low-risk patients. At a cut-off of 1, it yielded a sensitivity of 98.5% and an NPV of 98.17%, with an overall AUC of 0.868 (95% CI: 0.820–0.916). These results support the TIMI’s potential application for identifying low-risk ED patients.
Our findings showed an overall lower performance of TIMI compared to HEART in terms of AUC. This aligns with prior studies comparing the two systems. Poldervaart et al. (2017), in a large ED-based prospective cohort, found that the TIMI score had an AUC of 0.80 (95% CI: 0.78–0.83), which was lower than the HEART score’s AUC of 0.86 (95% CI: 0.84–0.88) [31]. Similarly, Ke et al. (2021), in a comprehensive meta-analysis, reported a pooled AUC of 0.80 (95% CI: 0.76–0.83) for TIMI, also lower than the HEART score’s pooled AUC of 0.87 [32].
Regarding safety thresholds, Poldervaart et al. demonstrated that a TIMI score of 0 classified 25.1% of patients as low-risk, with a MACE rate of 3.2% and an NPV of 97% (95% CI: 95–98%), closely matching the outcomes observed in our study. Ke et al. also reported a high pooled sensitivity (95%) and NPV for TIMI at low-risk thresholds, supporting its effectiveness in ruling out MACE.
In conclusion, although the TIMI score at a cut-off of 1 remains a highly sensitive tool for excluding MACE, its limited specificity limits its utility for nuanced decision-making. Nevertheless, in environments where expedited discharge of low-risk patients is essential, using TIMI = 1 as a threshold may offer a safe and practical approach when supported by clinical judgment.
GRACE score
In our cohort, GRACE showed moderate accuracy for short-term MACE prediction with an AUC of 0.815 (95% CI: 0.750–0.880), sensitivity of 91.9%, and NPV of 95.3% at a cut-off score of 91. This cut-off differs notably from thresholds commonly cited in literature, where higher values (e.g., 109 or 118) are often used to identify low-risk patients. The higher AUC and lower optimal cut-off in our study may reflect differences in population risk profiles due to laboratory data limitations.
Published evidence from large meta-analyses supports more conservative cut-off values. Kabiri et al. (2023) reported a pooled AUC of 0.71 (95% CI: 0.67–0.75) for GRACE at a cut-off of ≤ 100, with a high sensitivity of 0.96 (95% CI: 0.90–0.98) and a low specificity of 0.26 (95% CI: 0.16–0.40) [33]. Similarly, Ke et al. (2021) reported a pooled AUC of 0.70 (95% CI: 0.66–0.74), with a sensitivity of 0.78 (95% CI: 0.64–0.87) and specificity of 0.56 (95% CI: 0.46–0.66), emphasizing modest overall discriminatory performance [32]. These findings are consistent with previous literature suggesting that GRACE performs adequately for inpatient risk stratification but may be less suitable for early ED decision-making.
The relatively lower cut-off and higher AUC seen in our data likely reflect selection bias due to missing lab data, which restricted analysis to higher-risk patients. In conclusion, while the GRACE score demonstrated reasonable predictive value in our analysis, especially at lower thresholds, its practical use in emergency settings is constrained by its reliance on laboratory data and relatively lower performance compared to HEART and TIMI. Thus, GRACE may be better suited for inpatient risk stratification rather than front-line ED triage.
EDACS-ADP algorithm
In our study, EDACS-ADP demonstrated moderate predictive performance for identifying MACE within 6 weeks. The algorithm yielded an AUC of 0.803 (95% CI: 0.747–0.859), with a sensitivity of 91.2%, specificity of 69.4%, and an NPV of 95.98% at high-risk classification. This performance enabled the algorithm to correctly identify 91.2% of patients who experienced MACE, supporting its potential as an initial triage tool in ED settings.
While our results align with prior studies in terms of AUC, our sensitivity and NPV were noticeably lower. Wang et al. (2023) conducted a systematic review and reported a pooled sensitivity of 97% (95% CI: 0.95–0.99), specificity of 58% (95% CI: 0.53–0.63), and a pooled NPV exceeding 99% in low-risk patients. Their summary receiver operating characteristic (SROC) analysis yielded an AUC of 0.78 (95% CI: 0.74–0.81) [34]. similarly, In the validation study of Chae et al. (2022), EDACS-ADP showed a sensitivity of 99.5% (95% CI: 97.4–100%), classifying 31.3% (95% CI: 28.2–34.6%) of patients as low-risk, when unstable angina was excluded from the MACE definition [35].
In contrast, our study’s NPV of 95.98% falls below the commonly accepted > 98% safety threshold for early discharge in ED settings. This discrepancy may be due to the higher MACE prevalence in our cohort (24.8% versus 12% in Wang’s study), which directly lowers NPV, even when sensitivity remains high.
Although EDACS-ADP demonstrates an effective combination of sensitivity and specificity—especially in identifying low-risk patients—it may be less reliable in higher-acuity or referral-based EDs. Our findings suggest cautious application in such settings until further validation is achieved.
Strengths and limitations
This study evaluated five chest pain risk scores within a high-acuity emergency department setting, enhancing both statistical power and generalizability. The large sample size and elevated MACE incidence enabled robust ROC curve comparisons. Inclusion of Iranian ED populations—often underrepresented in validation literature—adds regional relevance and practical implications for triage in similar healthcare systems.
The study cohort spanned a broad risk spectrum, including self-referred low-risk patients and high-risk referrals, contributing to a pronounced bimodal distribution. This may have strengthened discriminative performance of scores that incorporate troponin and ECG parameters.
Nonetheless, limitations exist. The retrospective design limits control over confounding and introduces potential misclassification. Missing creatinine in 14.3% of cases affected GRACE score calculation, despite consistent imputation results. A transition between troponin assays may also influence score performance. Referral and spectrum bias are considerations due to the tertiary care setting. ECGs were interpreted by blinded experts, while clinical history collection by supervised interns may introduce variability.
Though findings are promising, external validation is needed before broader implementation, particularly in lower-acuity or resource-constrained environments.
Conclusion
The HEART and TIMI scores demonstrated the highest diagnostic accuracy for predicting 6-week major adverse cardiac events (MACE), particularly at thresholds of ≤ 3 and ≤ 1, respectively. These tools appear well-suited to guide early discharge decisions in emergency department (ED) settings, offering a favorable balance between sensitivity and negative predictive value (NPV). While the HET score yielded comparable performance, its clinical utility remains provisional pending further validation. In contrast, GRACE and EDACS-ADP exhibited moderate predictive capacity but were constrained by data availability and lower NPVs within this high-risk, referral-based cohort.
These findings underscore the importance of local validation, especially in tertiary care environments where elevated patient acuity and referral bias may distort the performance of risk stratification tools. Although our results affirm the potential of established scoring systems to support diagnostic decision-making in EDs, caution is warranted when extrapolating to broader clinical contexts. The influence of spectrum bias and the specialized nature of the study population may limit generalizability.
Notably, this study represents the first comprehensive validation of multiple cardiac risk scores within an Iranian emergency care setting. By aligning diagnostic performance with regional clinical workflows, our findings offer context-specific guidance and contribute to safer, evidence-based discharge planning.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgements
The authors wish to thank the emergency department staff working at Chamran and Al-Zahra hospitals. We also acknowledge the Trauma Data Registration Center affiliated with Isfahan University of Medical Sciences.
Abbreviations
- HEART
History, ECG, Age, Risk factors, Troponin
- TIMI
Thrombolysis in Myocardial Infarction
- GRACE
Global Registry of Acute Coronary Events
- EDACS
Emergency Department Assessment of Chest Pain Score
- HET
History, ECG, Troponin
- MACE
Major Adverse Cardiac Events
- STEMI
ST-Elevation Myocardial Infarction
- NSTEMI
Non-ST-Elevation Myocardial Infarction
- AUC
Area Under the Curve
- CI
Confidence Interval
- NRI
Net Reclassification Index
Author contributions
M.N.I., and N.S. contributed to the conception and design of the work. M.N.I., E.N.N. and H.R.M. contributed to data interpretation, drafting, and critical revision of the paper. H.R.M. helped with data collection and provided the initial draft of the manuscript. All authors read and approved the final version of the article.
Funding
This study was not funded by any organization. The authors wish to thank Vice Chancellery for research at Isfahan University of Medical Sciences for providing support for this research with project No. 3403168.
Data availability
Data are available upon reasonable request from the corresponding author.
Declarations
Ethics approval and consent to participate
This study was conducted in accordance with the Declaration of Helsinki (World Medical Association [WMA], 2013) and was approved by the Research Ethics Committee of Isfahan University of Medical Sciences (Approval No: IR.MUI.NUREMA.REC.1403.145). The ethics committee approved the use of anonymized data without requiring patient consent, as the data did not permit individual identification. Additionally, legislation on data safety and confidentiality was strictly followed.
Competing interests
The authors declare no competing interests.
Footnotes
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Contributor Information
Mehdi Nasr Isfahani, Email: m_nasr54@med.mui.ac.ir.
Hamidreza Mohseni, Email: itr@mui.ac.ir.
References
- 1.Zheng W, Wang G, Ma J, Wu S, Zhang H, Zheng J, Xu F, Wang J, Chen Y. Evaluation and comparison of six GRACE models for the stratification of undifferentiated chest pain in the emergency department. BMC Cardiovasc Disord. 2020;20(1):199. 10.1186/s12872-020-01476-3. PMID: 32334528; PMCID: PMC7183650. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Bhuiya FA, Pitts SR, McCaig LF. Emergency department visits for chest pain and abdominal pain: United States, 1999–2008. NCHS Data Brief. 2010;(43):1–8. PMID: 20854746. [PubMed]
- 3.Skinner JS, Smeeth L, Kendall JM, Adams PC, Timmis A, Chest pain guideline development group. NICE guidance. Chest pain of recent onset: assessment and diagnosis of recent onset chest pain or discomfort of suspected cardiac origin. Heart. 2010;96(12):974–8. 10.1136/hrt.2009.190066. PMID: 20538674. [DOI] [PubMed]
- 4.Schwartz GG, Fayyad R, Szarek M, DeMicco D, Olsson AG. Early, intensive statin treatment reduces ‘hard’ cardiovascular outcomes after acute coronary syndrome. Eur J Prev Cardiol. 2017;24(12):1294–6. Epub 2017 May 15. PMID: 28504565. [DOI] [PubMed] [Google Scholar]
- 5.Benjamin EJ, Muntner P, Alonso A, Bittencourt MS, Callaway CW, Carson AP, et al. Heart disease and stroke statistics-2019 update: a report from the American heart association. Circulation. 2019;139(10):e56–e528. 10.1161/CIR.0000000000000659. Erratum in: Circulation. 2020; 141(2):e33. PMID: 30700139. [DOI] [PubMed]
- 6.World Health Organization. The top 10 causes of death [Internet]. Geneva: WHO. 2024 [cited 2025 May 28]. Available from: WHO website.
- 7.Sun BC, Hsia RY, Weiss RE, Zingmond D, Liang LJ, Han W, et al. Effect of emergency department crowding on outcomes of admitted patients. Ann Emerg Med. 2013;61(6):605–e116. 10.1016/j.annemergmed.2012.10.026. Epub 2012 Dec 6. PMID: 23218508; PMCID: PMC3690784. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Januzzi J, McCarthy C. Evaluating chest pain in the emergency department: searching for the optimal gatekeeper. J Am Coll Cardiol. 2018;71(6):617–9. 10.1016/j.jacc.2017.11.065. [DOI] [PubMed] [Google Scholar]
- 9.Carlton EW, Then M, Cullen L, Khattab A, Greaves K, et al. Chest pain typicality’ in suspected acute coronary syndromes and the impact of clinical experience. Am J Med. 2015;128(10):1109–e162. Epub 2015 Apr 23. PMID: 25912206. [DOI] [PubMed] [Google Scholar]
- 10.Al-Zaiti S, Faramand Z, Alrawashdeh M, Sereika S, Martin-Gill C, et al. Comparison of clinical risk scores for triaging high-risk chest pain patients at the emergency department. Am J Emerg Med. 2019;37:461–7. 10.1016/j.ajem.2018.06.020. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Six AJ, Backus BE, Kelder JC. Chest pain in the emergency room: value of the HEART score. Neth Heart J. 2008;16(6):191–6. PMID: 18665203; PMCID: PMC2442661. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Backus BE, Six AJ, Kelder JC, Mast TP, van den Akker F, Mast EG, et al. Chest pain in the emergency room: a multicenter validation of the HEART Score. Crit Pathw Cardiol. 2010;9(3):164–9. 10.1097/HPC.0b013e3181ec36d8. PMID: 20802272. [DOI] [PubMed]
- 13.Antman EM, Cohen M, Bernink PJ, McCabe CH, Horacek T, Papuchis G, et al. The TIMI risk score for unstable angina/non-ST elevation MI: a method for prognostication and therapeutic decision making. JAMA. 2000;284(7):835–42. 10.1001/jama.284.7.835. PMID: 10938172. [DOI] [PubMed]
- 14.Granger CB, Goldberg RJ, Dabbous O, Pieper KS, Eagle KA, Cannon CP, et al. Predictors of hospital mortality in the global registry of acute coronary events. Arch Intern Med. 2003;163(19):2345–53. 10.1001/archinte.163.19.2345. PMID: 14581255. [DOI] [PubMed]
- 15.Than M, Flaws D, Sanders S, Doust J, Glasziou P, Kline J, et al. Development and validation of the emergency department assessment of chest pain score and 2 h accelerated diagnostic protocol. Emerg Med Australas. 2014;26(1):34–44. 10.1111/1742-6723.12164. Epub 2014 Jan 15. PMID: 24428678. [DOI] [PubMed]
- 16.Neumann JT, Twerenbold R, Ojeda F, Sörensen NA, Chapman AR, Shah ASV, et al. Application of high-sensitivity troponin in suspected myocardial infarction. N Engl J Med. 2019;380(26):2529–40. 10.1056/NEJMoa1803377. Erratum in: N Engl J Med. 2022;387(18):1724. PMID: 31242362. [DOI] [PubMed]
- 17.Löfmark H, Muhrbeck J, Eggers KM, Linder R, Ljung L, Martinsson A, et al. HEART-score can be simplified without loss of discriminatory power in patients with chest pain - introducing the HET-score. Am J Emerg Med. 2023; 74:104–11.10.1016/j.ajem.2023.09.037. Epub 2023 Oct 1. PMID: 37804822. [DOI] [PubMed]
- 18.Bozdereli Berikol G, Aydın H, Doğan H. Early discharging patients with chest pain using EDACS-ADP and COMPASS-MI risk predictors. Heart Vessels. 2022;37(8):1316–25. 10.1007/s00380-022-02036-9. Epub 2022 Feb 8. PMID: 35133498; PMCID: PMC8850102. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Chan Pin Yin D, Azzahhafi J, James S. Risk assessment using risk scores in patients with acute coronary syndrome. J Clin Med. 2020;9(9):3039. 10.3390/jcm9093039. PMID: 32967247; PMCID: PMC7565031. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Dinesh MV, Ramesh AC, Rangaswamy KMM, Veerendranath HK. A prospective risk stratification and validation of HEART, GRACE, and TIMI scores for chest pain patients presenting to the emergency department. J Emerg Pract Trauma. 2022;8(2):134–40. 10.34172/jept.2022.16. [Google Scholar]
- 21.Loten C, Isbister G, Jamcotchian M, Hullick C, MacElduff P, Attia J, et al. Adverse outcomes following emergency department discharge of patients with possible acute coronary syndrome. Emerg Med Australas. 2009;21(6):455–64. 10.1111/j.1742-6723.2009.01229.x. PMID: 20002715. [DOI] [PubMed]
- 22.Miller CD, Lindsell CJ, Khandelwal S, Chandra A, Pollack CV, Tiffany BR, et al. Is the initial diagnostic impression of noncardiac chest pain adequate to exclude cardiac disease? Ann Emerg Med. 2004;44:565–74. 10.1016/j.annemergmed.2004.03.021. [DOI] [PubMed] [Google Scholar]
- 23.Montassier E, Batard E, Gueffet JP, Trewick D, Le Conte P, et al. Outcome of chest pain patients discharged from a French emergency department: a 60-day prospective study. J Emerg Med. 2012;42(3):341–4. Epub 2011 Jan 19. PMID: 21247726. [DOI] [PubMed] [Google Scholar]
- 24.Ho YJ, Chen CH, Sung CW, et al. Risk stratification of patients with chest pain who have an unscheduled revisit to the emergency department. Int J Cardiol. 2023;383:96–101. 10.1016/j.ijcard.2023.04.043. [DOI] [PubMed] [Google Scholar]
- 25.Saadatagah S, Uddin MM, Weeks LD, et al. Clonal hematopoiesis risk score and all-cause and cardiovascular mortality in older adults. JAMA Netw Open. 2024;7(1):e2351927. 10.1001/jamanetworkopen.2023.51927. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Byrne C, Toarta C, Backus B, Holt T. The HEART score in predicting major adverse cardiac events in patients presenting to the emergency department with possible acute coronary syndrome: protocol for a systematic review and meta-analysis. Syst Rev. 2018;7(1):148. 10.1186/s13643-018-0816-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Fernando SM, Tran A, Cheng W, Rochwerg B, Taljaard M, Thiruganasambandamoorthy V, et al. Prognostic accuracy of the HEART score for prediction of major adverse cardiac events in patients presenting with chest pain: a systematic review and meta-analysis. Acad Emerg Med. 2019;26(2):140–51. 10.1111/acem.13649. [DOI] [PubMed] [Google Scholar]
- 28.Six AJ, Backus BE, Kelder JC. Chest pain in the emergency room: value of the HEART score. Neth Heart J. 2008;16(6):191–6. 10.1007/BF03086144. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Fernando S, Tran A, Cheng W, Rochwerg B, Taljaard M, Thiruganasambandamoorthy V, et al. Prognostic accuracy of the HEART score for prediction of major adverse cardiac events in patients presenting with chest pain: a systematic review and meta-analysis. Acad Emerg Med. 2018;26:140–51. 10.1111/acem.13649. [DOI] [PubMed] [Google Scholar]
- 30.Van Den Berg P, Body R. The HEART score for early rule out of acute coronary syndromes in the emergency department: a systematic review and meta-analysis. Eur Heart J Acute Cardiovasc Care. 2018;7:111–9. 10.1177/2048872617710788. [DOI] [PubMed] [Google Scholar]
- 31.Poldervaart J, Langedijk M, Backus B, Dekker I, Six A, Doevendans P, et al. Comparison of the GRACE, HEART, and TIMI score to predict major adverse cardiac events in chest pain patients at the emergency department. Int J Cardiol. 2017;227:656–61. 10.1016/j.ijcard.2016.10.080. [DOI] [PubMed] [Google Scholar]
- 32.Ke J, Chen Y, Wang X, Wu Z, Chen F. Indirect comparison of TIMI, HEART, and GRACE for predicting major cardiovascular events in patients admitted to the emergency department with acute chest pain: a systematic review and meta-analysis. BMJ Open. 2021;11:e048356. 10.1136/bmjopen-2020-048356. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Kabiri A, Gharin P, Forouzannia S, Ahmadzadeh K, Miri R, Yousefifard M, et al. HEART versus GRACE score in predicting the outcomes of patients with acute coronary syndrome: a systematic review and meta-analysis. Arch Acad Emerg Med. 2023;11(1). Available from: DOAJ. [DOI] [PMC free article] [PubMed]
- 34.Wang M, Hu Z, Miao L, Shi M, Gao Q. A systematic review of the applicability of emergency department assessment of chest pain score-accelerated diagnostic protocol for risk stratification of patients with chest pain. Clin Cardiol. 2023;46:1303–9. 10.1002/clc.24126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Chae B, Ahn S, Ryoo S, Kim Y, Seo D, Sohn C, et al. Performance analysis considering endpoints for three accelerated diagnostic protocols for chest pain. Am J Emerg Med. 2022;64:51–6. 10.1016/j.ajem.2022.11.020. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
Data are available upon reasonable request from the corresponding author.