Archives of Academic Emergency Medicine. 2023; 11(1): e50 REV I EW ART I C L E HEART versus GRACE Score in Predicting the Outcomes of Patients with Acute Coronary Syndrome; a Systematic Re- view and Meta-Analysis Ali Kabiri1, Pantea Gharin1, Seyed Ali Forouzannia2, Koohyar Ahmadzadeh1, Reza Miri3∗, Mahmoud Yousefifard3,1 † 1. Physiology Research Center, Iran University of Medical Sciences, Tehran, Iran. 2. Department of Medicine, School of Medicine, Shahid Beheshti University of Medical Sciences, Tehran, Iran. 3. Prevention of Cardiovascular Disease Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran Received: May 2023; Accepted: June 2023; Published online: 19 July 2023 Abstract: Introduction: Several scoring systems have been proposed to predict the outcomes of patients with ischemic heart disease. Global Registry of Acute Coronary Events (GRACE) and History, ECG, Age, Risk Factors, and Troponin (HEART) scores are two of the more widely used risk prediction tools in patients with acute coronary syndrome (ACS). The present systematic review and meta-analysis aimed to compare the value of GRACE and HEART scores in the outcome prediction of ACS patient. Methods: The online databases of Medline, Embase, Web of Science, and Scopus were search until September 2022 for articles directly comparing GRACE and HEART scores value in prediction of outcome in patients with ACS. GRACE score cut-offs were categorized into two groups of less than and equal to 100 and more than 100, and HEART score cut-offs were categorized into three groups of less than 4, equal to 4, and more than 4. Investigated outcomes were major adverse cardiovascular events (MACE), acute myocardial infraction (AMI) and all-cause mortality. Results: 25 articles were included. The sensitivity and specificity of the GRACE score for prediction of MACE were 0.96 and 0.26 for cut-offs of ≤ 100, and 0.58 and 0.69 for cut-offs of >100, respectively. The sensitivity and specificity of the HEART score for prediction of MACE were 0.99 and 0.16 for cut-offs less than 4, 0.93 and 0.47 for equal to 4, and 0.77 and 0.78 for cut-offs greater than 4. GRACE score was shown to be predictive of AMI with sensitivity and specificity of 0.95 and 0.29, respectively. The analysis for the value of HEART score in the prediction of AMI a sensitivity and specificity of 0.94 and 0.48, respectively. The risk scores were not found to be suitable predictors of all-cause mortality. Conclusion: The results demonstrated the low specificity of GRACE and HEART scores in predicting the MACE, AMI and all-cause mortality, irrespective of the utilized cut-off. Considering the acceptable sensitivity of two scores in predicting the MACE and AMI, these scores were more suitable to be used as a rule-out tool for identification of ACS patients with low risk of developing adverse outcomes. Keywords: Acute coronary syndrome; risk assessment; sensitivity; specificity; decision tools Cite this article as: Kabiri A, Gharin P, Forouzannia SA, Ahmadzadeh K, Miri R, Yousefifard M. 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): e50. https://doi.org/10.22037/aaem.v11i1.2001. ∗Corresponding Author: Reza Miri; Prevention of Cardiovascular Dis- ease Research Center, Imam Hossein Hospital, Madani Avenue, Tehran, Iran, Phone/Fax: +982177582721; Email: dr.rezamiri1@gmail.com, ORCID: https://orcid.org/0000-0002-8568-9948. † Corresponding Author: Mahmoud Yousefifard; Physiology Research Center, Iran University of Medical Sciences, Hemmat Highway, P.O Box: 14665-354, Tehran, Iran; Phone/Fax: +982186704771; Email: yousefifard20@gmail.com / yousefifard.m@iums.ac.ir, ORCID: https://orcid.org/0000-0001-5181-4985. 1. Introduction Ischemic heart disease (IHD) is the most common cardiovas- cular disease and accounts for a vast amount of cardiovas- cular disease burden. It has been speculated that IHD was the cause of 9.14 million deaths in 2019 (1). Knowledge of the outcome of cardiovascular diseases can aid in the timely identification of high-risk patients and their management (2). Several scoring tools and markers have been proposed to predict the outcomes of cardiovascular diseases, especially mortality and major adverse cardiovascular events (MACE) (3-5). Global Registry of Acute Coronary Events (GRACE) This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index A. Kabiri et al. 2 and History, ECG, Age, Risk Factors, and Troponin (HEART) scores are two of the more widely used risk prediction tools in patients with acute coronary syndrome (ACS). The GRACE score was introduced in 2007 and is calculated based on patient’s age, killip class, heart rate, systolic blood pressure, ST segment changes, creatinine levels, and elevated cardiac markers to assess the risk of unfavorable outcomes in ACS patients (6). The HEART risk score was introduced in 2008 for identification of ACS patients with a higher risk of adverse outcomes (7). The predictive value of these tools has been investigated in ACS patients with promising results (2, 7-11). In a recent systematic review and meta-analysis, Ke et al. (12) have indirectly evaluated GRACE and HEART scores in the prediction of MACE in ACS patients. Their results have demonstrated that HEART and GRACE scores could predict MACE with a sensitivity of 96% and 78% and specificity of 50% and 56%, respectively, and have concluded that HEART is more accurate than GRACE in the prediction of MACE in ACS patients. Considering that direct comparisons are preferred as a basis for drawing conclusions, the present systematic review and meta-analysis was designed with the aim of directly compar- ing the value of GRACE and HEART scores in the prediction of ACS patient outcomes. 2. Methods 2.1. Study design and search strategy This systematic review and meta-analysis was designed to evaluate and compare the value of HEART and GRACE tools in the prediction of ACS patient outcomes. P (patients): acute coronary syndrome patients, I (Index test): GRACE and HEART tools, C (Comparison): ACS patients without the outcome of the study, O (Outcome): Major adverse cardiovascular events (MACE), Acute myocar- dial infarction (AMI), and all-cause mortality were chosen as the definition of PICO for the current review. MeSH terms of PubMed and Emtree terms of Embase databases were used to acquire related keywords. Chosen keywords were further tailored for the aim of this study by reviewing relevant arti- cles and consultation with experts in the field. The online databases of PubMed, Embase, Web of Science, and Scopus were searched until September 10th, 2022, with the specific search strategies devised for each database (Supplementary material 1). A manual review of Google and Google scholar search engines and references of relevant articles was also performed to access any possibly missed studies. 2.2. Selection criteria All articles with direct comparisons of GRACE and HEART scores in the adult population were included. The exclusion criteria of this study were review articles, non-English arti- cles, articles with a design of indirect comparison, case se- ries, duplicate records, and articles with no report of required data for this review. 2.3. Data extraction The title and abstract of the retrieved records from online databases were screened by two independent reviewers and after the full-text screening, relevant studies were included. Reported data were extracted into a checklist designed ac- cording to PRISMA guidelines. The information in the check- list comprised study characteristics (first author, year, coun- try), sample size, age, and male number, the outcome of the study and its definition, event rate of the whole population, and the number of patients who developed outcome in each group, follow-up duration, sensitivity, specificity, true and false positives, and true and false negatives. Any conflicts of opinion were resolved by consulting a third reviewer. 2.4. Quality assessment and Certainty of evi- dence The quality of the included articles was evaluated using the Quality Assessment of Prognostic Accuracy Studies (QUA- PAS) risk of bias assessment tool (13). According to this tool, studies are evaluated in two sections of risk of bias (consist- ing of patient selection, index test, reference standard, flow and timing and analysis) and applicability (consisting of pa- tient selection, index test, and reference standard). The certainty of evidence of the included articles was evalu- ated by the Grades of Recommendation, Assessment, Devel- opment, and Evaluation (GRADE) guidelines (14). 2.5. Statistical analysis STATA 17.0 statistical software was used to perform the anal- ysis. Reported data were recorded as true and false positives and true and false negatives. “midas” package was used to analyze the data. Findings were reported as pooled sensi- tivity, pooled specificity, area under the curve (AUC), posi- tive and negative likelihood ratios, and diagnostic odds ra- tio. Since the results of the articles were reported by different cut-offs, we stratified the analysis based on the reported cut- offs. For this purpose, we categorized GRACE score cut-offs into two groups of less than and equal to 100 and more than 100, and HEART score cut-offs into three groups of less than 4, equal to 4, and more than 4. The publication bias of the in- cluded studies was assessed using Deeks’ asymmetry funnel plot. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 3 Archives of Academic Emergency Medicine. 2023; 11(1): e50 3. Results 3.1. Study flow and characteristics The systematic search in online databases yielded 1860 non- duplicate records. 59 reports were assessed for eligibility after the title and abstract screening and 25 articles were chosen to be included in this study (2, 15-38). 3 articles were retrieved by manual search, none of which were included (Figure 1). The articles comprised data on 21389 suspected ACS patients (55.01% male), with a follow-up of at least 30 days (15 articles) with a maximum follow-up time of roughly 8 years. The as- sessed outcomes were MACE (18 articles), AMI (6 articles), and all-cause of mortality (3 articles). 19 studies were de- signed as prospective cohorts, and 6 were retrospective co- horts. The characteristics of the included articles are demon- strated in table 1. 3.2. Value of GRACE and HEART risk scores in the prediction of MACE Studies included 17697 patients (56.1% male) with an event rate of 18.54% (3282 patients) in the evaluation of the value of GRACE and HEART scores in the prediction of MACE. The results of the analysis for GRACE score with cut-offs of less than and equal to 100 and more than 100 showed AUCs of 0.71 (95% CI: 0.67, 0.75) and 0.66 (95% CI: 0.62, 0.70), re- spectively. The sensitivity and specificity of the GRACE score were 0.96 (95% CI: 0.90, 0.98) and 0.26 (95% CI: 0.16, 0.40) for cut-offs of ≤ 100 and 0.58 (95% CI: 0.53, 0.64) and 0.69 (0.61, 0.77) for cut-offs of >100, respectively. The AUC of HEART scores less than 4 and equal to 4 was cal- culated as 0.98 (95% CI: 0.96, 0.99) and 0.72 (95% CI: 0.68, 0.76), respectively. The AUC of HEART scores more than 4 was 0.84 (95% CI: 0.81, 0.87). The sensitivity and specificity of the HEART scores were 0.99 and 0.16 for scores less than 4, 0.93 and 0.47 for equal to 4, and 0.77 and 0.78 for scores greater than 4. Based on the presented data, the HEART score is best predictive of MACE with a cut-off score of 4 (Table 2). Although it should be mentioned that only 3 articles had data on the prognostic value of HEART score with cut-offs of less than 4 (Figure 2 and Table 2). 3.3. Value of GRACE and HEART risk scores in the prediction of AMI Studies comprised 3591 patients (55.4% male) with an event rate of 13.23% (475 patients) in the assessment of the prog- nostic values of GRACE and HEART for the prediction of AMI. AUC of the GRACE score for the prediction of AMI was 0.72 (95% CI: 0.68, 0.76) and the sensitivity and specificity were 0.95 (95% CI: 0.86, 0.98) and 0.29 (95% CI: 0.16, 0.46), respec- tively (Figure 3 and Table 2). All studies, except two (21, 23), utilized cut-offs of less than 100. A sensitivity analysis was performed for reports with cut-offs of less than 100 and the results demonstrated an AUC of 0.76 (95% CI: 0.72, 0.80), sen- sitivity of 0.98 (95% CI: 0.93, 0.99) and specificity of 0.18 (95% CI: 0.09, 0.31) for the value of GRACE score in prediction of AMI. The analysis for the value of HEART score in the prediction of AMI revealed an AUC of 0.86 (95% CI: 0.82, 0.88) and a sensi- tivity and specificity of 0.94 (95% CI: 0.88 and 0.97) and 0.48 (95% CI: 0.32, 0.64), respectively. One article had utilized a cut-off of 3 (16), two articles had not reported the utilized cut-off (21, 24) and the remaining three articles had utilized a cut-off of 4. A sensitivity analysis was performed for the three articles utilizing a cut-off of 4 and the results showed an AUC of 0.83 (95% CI: 0.80, 0.86), a sensitivity of 0.94 (95% CI: 0.86, 0.97) and a specificity of 0.51 (95% CI: 0.36, 0.66). 3.4. Value of GRACE and HEART risk scores in the prediction of all-cause mortality 1903 patients (79.8% male, 15.24% event rate) were included in the meta-analysis for the value of GRACE and HEART scores in the prediction of all-cause mortality. The results of the analysis showed an AUC of 0.75 (95% CI: 0.71, 0.79) for GRACE and 0.65 (95% CI: 0.61, 0.69) for the HEART score. The sensitivity and specificity of GRACE were 0.82 (95% CI: 0.75, 0.87) and 0.51 (95% CI: 0.42, 0.61), respectively and the sensi- tivity and specificity of HEART score were 0.78 (95% CI: 0.57, 0.90) and 0.56 (95% CI: 0.49, 0.62). Two studies utilized a cut- off of more than 100 for the GRACE score and one study uti- lized a cut-off of 4 for the HEART score, the remaining studies had not reported the cut-off used. Due to the scarce number of included studies, sensitivity analysis was not performed (Figure 4, Table 2). 3.5. Publication bias Publication bias was only assessed for the outcome of MACE, which had more than 10 included studies. No publication bias was observed in the evaluation of the value of GRACE and HEART scores in the prediction of MACE (Figure 5). 3.6. Risk of bias assessment The risk of bias in the domain of patient selection was eval- uated to be unclear in eight studies, due to no mention of their sampling method, and high in three studies due to con- venient sampling. Eight studies had not mentioned the cri- teria for choosing the risk scores cut-offs and were rated as unclear in risk of bias in the domain of index test and four studies were rated as high in risk of bias in this domain due to choosing their cut-offs based on the calculated AUCs. Out- come assessment protocol was unclear in seven articles and one article did not provide any outcome definition and was rated as high in risk of bias in the domain of outcome. Three studies were rated as unclear in risk of bias in the domain of flow and timing due to possible loss to follow-ups. Two stud- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index A. Kabiri et al. 4 ies were rated as unclear in the risk of bias in the domain of analysis due to having possible competing events. One study was rated as high in the applicability of outcome due to no outcome definition. The articles were rated as low in the re- mainder of the domains. Overall, the included studies were judged to have a serious risk of bias (Table 3). 3.7. Certainty of evidence The certainty of the evidence of the included studies was as- sessed using GRADE guidelines. The included studies were designed as cohort studies and according to GRADE guide- lines, the base level of evidence was set as high for compara- tive test accuracy studies. The level of evidence for the outcome of MACE was rated down one score due to the serious risk of bias in included ar- ticles and thus both scores had a moderate level of evidence for the outcome of MACE. The level of evidence in the out- comes of AMI and all-cause mortality was rated down two scores for serious risk of bias, and the inability of assessing publication bias due to the limited included studies. The level of evidence for outcomes of AMI and all-cause mortality was rated as low for both risk scores (Table 4). 4. Discussion The current study is the first meta-analysis conducted on the direct comparison of the value of GRACE and HEART scores for the prediction of adverse outcomes in ACS patients. We evaluated the predictive value of GRACE and HEART scores in outcomes of MACE, AMI, and all-cause mortality and our results showed that utilizing GRACE and HEART scores with appropriate cut-offs can predict MACE and AMI with accept- able sensitivities. The included studies had utilized various cut-off values. Our analysis demonstrated that the predictive performance of the GRACE score for the outcome of MACE greatly improves when utilized with cut-off values less than 100 (Sensitivity of 0.96 for cut-off values less than 100 and sensitivity of 0.59 for cut-off values more than 100). The predictive value of the HEART score was shown to be higher when utilized with cut- offs of less than 4 (Sensitivity of 0.99), or equal to 4 (sensitiv- ity of 0.93), and our results suggest that utilization of HEART score with a cut-off value of more than 4 (sensitivity of 0.77) cannot be administered as a predictive tool for MACE. GRACE and HEART tools were also demonstrated to be good predictors of AMI, with sensitivities of 0.95 and 0.94 respec- tively. Further analysis revealed that utilization of the GRACE score with cut-off values of less than 100, slightly improves its predictive value for AMI (sensitivity of 0.98) while subgroup- ing the analysis of the HEART score by the cut-off value of 4, did not reveal any significant changes to its predictive value for AMI. In contrast to the acceptable performance of GRACE and HEART scores for the prediction of MACE and AMI, in our analysis, these scores were not found to be proper predictors of all-cause mortality (sensitivities of 0.82 and 0.78 respec- tively). Although it should be mentioned that our results are limited by the scarce number of included studies investigat- ing the outcome of all-cause mortality. Overall, our results have demonstrated low specificity for GRACE and HEART scores in all three outcomes, irrespective of the utilized cut- off. We believe this reiterates the fact that such scores are more suitable to be used as a rule-out tool for identifica- tion of ACS patients with low risk of developing adverse out- comes, rather than as a rule-in tool to identify patients with higher chances of developing adverse outcomes. Previous systematic reviews have evaluated the predictive value of GRACE and HEART scores separately or as an in- direct comparison. Van Den Berg et al, suggested that the HEART score could be predictive of MACE with a sensitiv- ity and specificity of 0.96 and 0.47 respectively (39). Whereas Ke et al (12) indirectly compared the predictive performance of GRACE and HEART scores and demonstrated a sensitiv- ity and specificity of 0.96 and 0.50 for GRACE and sensitivity and specificity of 0.78 and 0.56 for HEART scores in the pre- diction of MACE. Our results for the predictive performance of GRACE for MACE are in line with previous reviews, how- ever, our results demonstrated a better predictive value for the HEART score. Moreover, we suggest that GRACE and HEART scores can also be used to predict AMI in ACS pa- tients. Although the specified cut-off value chosen for the in- terpretation of these risk scores is important and can vastly affect their predictive capabilities. Considering the fairly similar predictive performances of GRACE and HEART scores in our review, the differences in their design should be kept in mind for better application of these scores. HEART score variables are more readily avail- able, and the overall score can be easily calculated. This score relies on the judgment of patient history suspicion which can be challenging due to no exact definition for high, moder- ate, or slight suspicion, however it has been argued that in- corporating clinical gestalt can improve the performance ca- pabilities of the score (26). GRACE score has been found to be more complex which might require a computer for calcu- lations. There are also fundamental differences in the aims of each score. While HEART score aims to identify low-risk chest pain patients for early discharge, GRACE score was de- rived for high-risk patients investigating the need for invasive therapy and not for evaluating individuals with undifferenti- ated chest pain (40-42). It should also be noted that GRACE and HEART scores should be used to enhance the decision- making process of the physician and they are never meant to replace clinical decision-making. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 5 Archives of Academic Emergency Medicine. 2023; 11(1): e50 5. Limitations This systematic review has a few limitations. First, the in- cluded studies had differing patient selection criteria. While some studies had included all suspected ACS patients, some had only investigated non-ST elevation myocardial infarc- tion (MI) ACS patients, and few studies had only included ST elevation MI ACS patients. Thus, although our analy- sis is indicative of all types of ACS patients, further studies could evaluate the value of GRACE and HEART scores in spe- cific ACS patient settings. The definition of MACE in the in- cluded articles could affect the predictive value of the GRACE and HEART scores. The articles had varying MACE defini- tions, ranging from a combined endpoint of death or MI to more composite endpoints, which made a subgroup analy- sis impossible. The included articles had not reported the treatment plan of their patients. Considering that different treatment regimens can have an effect on the outcome of patients, future studies could aim to evaluate the predictive value of scoring systems in populations receiving uniform treatments. 6. Conclusion The results demonstrated the low specificity of GRACE and HEART scores in predicting the MACE, AMI and all-cause mortality, irrespective of the utilized cut-off. Considering the acceptable sensitivity of two scores in predicting the MACE and AMI, these scores were more suitable to be used as a rule-out tool for identification of ACS patients with low risk of developing adverse outcomes. 7. Declarations 7.1. Acknowledgments None. 7.2. Conflict of interest The authors declare that they have no competing interests. 7.3. Funding and support This study was funded by Shahid Beheshti University of Med- ical Sciences. 7.4. Authors’ contribution Study design: MY, RM Data gathering: AK, PG, SAF Analysis: MY, KA Interpretation of results: all authors Drafting and revising: all authors All authors read and approved final version. 7.5. Using artificial intelligence chatbots None. 7.6. Availability of data and materials The gathered data and checklist can be provided to qualified researchers with the intent of replicating the procedure and results. References 1. Roth GA, Mensah GA, Johnson CO, Addolorato G, Ammi- rati E, Baddour LM, et al. Global burden of cardiovascu- lar diseases and risk factors, 1990–2019: update from the GBD 2019 study. Journal of the American College of Car- diology. 2020;76(25):2982-3021. 2. Huang Z, Wang K, Yang D, Gu Q, Wei Q, Yang Z, et al. The predictive value of the HEART and GRACE scores for major adverse cardiac events in patients with acute chest pain. Intern Emerg Med. 2021;16(1):193-200. 3. Ramezani F, Ahmadi S, Faridaalee G, Baratloo A, Yousefi- fard M. Value of Manchester acute coronary syndromes decision rule in the detection of acute coronary syn- drome; a systematic review and meta-analysis. Emer- gency. 2018;6(1). 4. Haji Aghajani M, Madani Neishaboori A, Ahmadzadeh K, Toloui A, Yousefifard M. The association between apolipoprotein A-1 plasma level and premature coro- nary artery disease: A systematic review and meta- analysis. International journal of clinical practice. 2021;75(11):e14578. 5. Aghajani MH, Toloui A, Ahmadzadeh K, Neishaboori AM, Yousefifard M. Premature Coronary Artery Disease and Plasma Levels of Interleukins; a Systematic Scoping Re- view and Meta-Analysis. Archives of Academic Emer- gency Medicine. 2022;10(1). 6. Tang EW, Wong C-K, Herbison P. Global Registry of Acute Coronary Events (GRACE) hospital discharge risk score accurately predicts long-term mortality post acute coro- nary syndrome. American heart journal. 2007;153(1):29- 35. 7. Laureano-Phillips J, Robinson RD, Aryal S, Blair S, Wilson D, Boyd K, et al. HEART Score Risk Stratification of Low- Risk Chest Pain Patients in the Emergency Department: A Systematic Review and Meta-Analysis. Annals of Emer- gency Medicine. 2019;74(2):187-203. 8. Khalil MH, Sekma A, Yaakoubi H, Bel Haj Ali K, Msolli MA, Beltaief K, et al. 30 day predicted outcome in undifferen- tiated chest pain: multicenter validation of the HEART score in Tunisian population. BMC Cardiovascular Dis- orders. 2021;21(1):1-7. 9. Soares III WE, Knee A, Gemme SR, Hambrecht R, Dybas S, Poronsky KE, et al. A prospective evaluation of clin- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index A. Kabiri et al. 6 ical HEART score agreement, accuracy, and adherence in emergency department chest pain patients. Annals of Emergency Medicine. 2021;78(2):231-41. 10. Rao K, De Silva K, Sood A, Denniss AR, Hsu C-J. Predicting Patients With Troponin Positive Chest Pain and Unob- structed Coronary Arteries With Electrocardiogram, Tro- ponin Kinetics and GRACE Score. Heart, Lung and Circu- lation. 2022;31(9):1219-27. 11. Cedro AV, Mota DM, Ohe LN, Timerman A, Costa JR, Cas- tro LdS. Association between Clinical Risk Score (Heart, Grace and TIMI) and Angiographic Complexity in Acute Coronary Syndrome without ST Segment Elevation. Ar- quivos Brasileiros de Cardiologia. 2021;117:281-7. 12. Ke J, Wang X, Wu Z, Chen F. Indirect comparison of TIMI, HEART and GRACE for predicting major cardiovascular events in patients admitted to the emergency depart- ment with acute chest pain: a systematic review and meta-analysis. BMJ open. 2021;11(8):e048356. 13. Lee J, Mulder F, Leeflang M, Wolff R, Whiting P, Bossuyt PM. QUAPAS: An Adaptation of the QUADAS-2 Tool to Assess Prognostic Accuracy Studies. Ann Intern Med. 2022;175(7):1010-8. 14. Yang B, Mustafa RA, Bossuyt PM, Brozek J, Hultcrantz M, Leeflang MM, et al. GRADE Guidance: 31. Assessing the certainty across a body of evidence for comparative test accuracy. Journal of clinical epidemiology. 2021;136:146- 56. 15. Al-Zaiti SS, Faramand Z, Alrawashdeh MO, Sereika SM, Martin-Gill C, Callaway C. Comparison of clin- ical risk scores for triaging high-risk chest pain pa- tients at the emergency department. Am J Emerg Med. 2019;37(3):461-7. 16. Carlton EW, Khattab A, Greaves K. Identifying Patients Suitable for Discharge After a Single-Presentation High- Sensitivity Troponin Result: A Comparison of Five Estab- lished Risk Scores and Two High-Sensitivity Assays. Ann Emerg Med. 2015;66(6):635-45 e1. 17. Chae MK, Kim EK, Jung KY, Shin TG, Sim MS, Jo IJ, et al. Triple rule-out computed tomography for risk stratifica- tion of patients with acute chest pain. J Cardiovasc Com- put Tomogr. 2016;10(4):291-300. 18. Chen XH, Jiang HL, Li YM, Chan CPY, Mo JR, Tian CW, et al. Prognostic values of 4 risk scores in Chinese pa- tients with chest pain: Prospective 2-centre cohort study. Medicine (Baltimore). 2016;95(52):e4778. 19. Chew PG, Frost F, Mullen L, Fisher M, Zadeh H, Grainger R, et al. A direct comparison of decision rules for early discharge of suspected acute coronary syndromes in the era of high sensitivity troponin. Eur Heart J Acute Cardio- vasc Care. 2019;8(5):421-31. 20. Dinesh MV, Ramesh AC, Rangaswamy KMM, Veeren- dranath HK. A prospective risk stratification and valida- tion of HEART, GRACE and TIMI scores for chest pain pa- tients presenting to the emergency department. Journal of Emergency Practice and Trauma. 2022;8(2):134-40. 21. Dupuy AM, Pasquier G, Thiebaut L, Roubille F, Sebbane M, Cristol JP. Additive value of bioclinical risk scores to high sensitivity troponins-only strategy in acute coro- nary syndrome. Clin Chim Acta. 2021;523:273-84. 22. Han K-W, Chen S-Y, Weng Y-M, Ng C-J, Chiu T-F, Hsieh IC, et al. Validation of different score systems in predict- ing cardiac arrest occurrence of ST-elevation myocardial infarction. Hong Kong Journal of Emergency Medicine. 2017;24(5):224-9. 23. Hrecko J, Dokoupil J, Pudil R. Comparison of six decision aid rules for diagnosis of acute myocardial infarction in elderly patients presenting to the emergency department with acute chest pain. Bratisl Lek Listy. 2022;123(4):282- 90. 24. Jukneviciene R, Juknevicius V, Jasiunas E, Rasciute B, Barysiene J, Mataciunas M, et al. Chest pain in the emergency department: From score to core- A prospective clinical study. Medicine (Baltimore). 2022;101(29):e29579. 25. Liu N, Sakamoto JT, Cao J, Koh ZX, Ho AFW, Lin Z, et al. Ensemble-Based Risk Scoring with Extreme Learning Machine for Prediction of Adverse Cardiac Events. Cog- nitive Computation. 2017;9(4):545-54. 26. Ng M, Tan HJG, Gao F, Tan JWC, Lim SH, Ong MEH, et al. Comparative prospective study of the performance of chest pain scores and clinical assessment in an emer- gency department cohort in Singapore. J Am Coll Emerg Physicians Open. 2020;1(5):723-9. 27. Poldervaart JM, Langedijk M, Backus BE, Dekker IMC, Six AJ, Doevendans PA, et al. Comparison of the GRACE, HEART and TIMI score to predict major adverse cardiac events in chest pain patients at the emergency depart- ment. Int J Cardiol. 2017;227:656-61. 28. Reaney PDW, Elliott HI, Noman A, Cooper JG. Risk strat- ifying chest pain patients in the emergency department using HEART, GRACE and TIMI scores, with a single con- temporary troponin result, to predict major adverse car- diac events. Emerg Med J. 2018;35(7):420-7. 29. Ruangsomboon O, Thirawattanasoot N, Chakorn T, Lim- suwat C, Monsomboon A, Praphruetkit N, et al. The util- ity of the 1-hour high-sensitivity cardiac troponin T algo- rithm compared with and combined with five early rule- out scores in high-acuity chest pain emergency patients. Int J Cardiol. 2021;322:23-8. 30. Sakamoto JT, Liu N, Koh ZX, Fung NX, Heldeweg ML, Ng JC, et al. Comparing HEART, TIMI, and GRACE scores for prediction of 30-day major adverse cardiac events in high acuity chest pain patients in the emergency department. Int J Cardiol. 2016;221:759-64. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 7 Archives of Academic Emergency Medicine. 2023; 11(1): e50 31. Shin YS, Ahn S, Kim YJ, Ryoo SM, Sohn CH, Kim WY. Risk stratification of patients with chest pain or anginal equiv- alents in the emergency department. Intern Emerg Med. 2020;15(2):319-26. 32. Singer AJ, Than MP, Smith S, McCullough P, Barrett TW, Birkhahn R, et al. Missed myocardial infarctions in ED patients prospectively categorized as low risk by estab- lished risk scores. Am J Emerg Med. 2017;35(5):704-9. 33. Steiro OT, Tjora HL, Langorgen J, Bjorneklett R, Nygard OK, Skadberg O, et al. Clinical risk scores identify more patients at risk for cardiovascular events within 30 days as compared to standard ACS risk criteria: the WESTCOR study. Eur Heart J Acute Cardiovasc Care. 2021;10(3):287- 301. 34. Tekin E, Ozlu I, Onur Can N, Yilmaz S. Comparison of the integrated pulmonary index with cardiac risk scores in acute coronary syndromes. Annals of Clinical and Ana- lytical Medicine. 2021;12(03). 35. Torralba F, Navarro A, la Hoz JC, Ortiz C, Botero A, Alar- con F, et al. HEART, TIMI, and GRACE Scores for Predic- tion of 30-Day Major Adverse Cardiovascular Events in the Era of High-Sensitivity Troponin. Arq Bras Cardiol. 2020;114(5):795-802. 36. Wong CP, Lui CT, Sung JG, Lam H, Fung HT, Yam PW. Prognosticating Clinical Prediction Scores Without Clini- cal Gestalt for Patients With Chest Pain in the Emergency Department. J Emerg Med. 2018;54(2):176-85. 37. Yang B, Bai L, Zhang Y, Cheng Y, Zhao C, Huang B, et al. The Value of Different Short-Term Risk Scoring Models in Predicting Long-Term Death of Acute Myocardial Infarc- tion. J Clin Med. 2022;11(17). 38. Zheng W, Wang G, Ma J, Wu S, Zhang H, Zheng J, et al. Evaluation and comparison of six GRACE mod- els for the stratification of undifferentiated chest pain in the emergency department. BMC Cardiovasc Disord. 2020;20(1):199. 39. Van Den Berg P, Body R. The HEART score for early rule out of acute coronary syndromes in the emergency de- partment: a systematic review and meta-analysis. Eur Heart J Acute Cardiovasc Care. 2018;7(2):111-9. 40. Six AJ, Backus BE, Kelder JC. Chest pain in the emergency room: value of the HEART dvotr. Netherlands Heart Jour- nal. 2008;16(6):191-6. 41. Backus BE, Six AJ, Kelder JH, Gibler WB, Moll FL, Doeven- dans PA. Risk Scores for Patients with Chest Pain: Evalu- ation in the Emergency Department. Current Cardiology Reviews. 2011;7(1):2-8. 42. Brady W, de Souza K. The HEART score: A guide to its application in the emergency department. Turk J Emerg Med. 2018;18(2):47-51. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index A. Kabiri et al. 8 Table 1: Characteristics of included studies Study Country Design Follow up (day) Sample size Male (n) Age* (year) Outcome: number GRACE cut-off HEART cut-off Al-Zaiti 2018 US RCS 30 750 433 59±17 MACE: 33 109 4 Carlton 2015 UK PCS 30 959 564 58±13.3 AMI: 79 60, 80 3, 4 Chae 2016 South Korea PCS 30 1024 594 58 (50-69) MACE: 126 108 4 Chen 2016 China PCS 30, 180 833 461 65.1±14.5 MACE: 121 109, 114 4, 5 Chew 2018 UK PCS 42, 365 1642 858 59.35±18.54 MACE: 279 AMI: 180 76 4 Dinesh 2022 India PCS 42 199 138 51.61±16.47 MACE: 76 119 7 Dupuy 2021 France PCS 30 160 94 73 to 80 AMI: 37 Death: 13 NR NR Han 2017 Taiwan RCS 180 249 203 61.7±14.91 Death: 41 121 NR Hrecko 2022 Finland RCS 30 250 126 78.5±8.2 AMI: 48 109 4 Huang 2021 China PCS 30 509 293 59.77±14.9 MACE: 92 106 4 Jukneviciene 2022 Lithuania PCS 180 146 95 63±13.4 AMI: 51 NR NR Liu 2017 Singapore PCS NR 797 542 NR MACE: 146 108 5 Mingwei NG 2020 Singapore PCS 30 1195 817 NR MACE: 135 109 4 Poldervaart 2017 Netherlands PCS 42 1748 937 62±14 MACE: 326 73 4 Reaney 2017 UK PCS 30 1000 574 62.4±15.6 MACE: 189 56, 119 4, 7 Ruangsomboon 2020 Thailand PCS 30 350 185 66.3±15.1 MACE: 59 56, 66 3, 4 Sakamoto 2016 Singapore PCS 30 604 417 60.8±13.2 MACE: 215 76, 111 4, 5 Shin 2020 Korea RCS 30, 90 1247 758 62±12.7 MACE: 211 109 4 Singer 2017 US PCS 30 434 251 56.64±11.15 AMI: 80 51 4 Steiro 2021 Norway RCS 30, 180 932 562 63±16.33 MACE: 191 90, 109 4 Tekin 2021 Turkey PCS 30 381 195 NR MACE: 105 115 6 Torralba 2020 Colombia PCS 30 519 291 64.31±12.11 MACE: 224 109 4 Wong 2017 Hong Kong PCS 30 1081 565 48±27 MACE: 164 50,75,100 1, 2 Yang 2022 China RCS 2956 1494 1221 63 (53-72) Death: 236 126 4 Zheng 2020 China PCS 30 2886 1447 64±13.5 MACE: 590 81 4 Data are presented as mean ± standard deviation or median (interquartile range). AMI: Acute myocardial infarction, MACE: Major adverse cardiovascular event, NR: Not reported, PCS: Prospective cohort study, RCS: Retrospective cohort study Table 2: Performance of GRACE and MACE scores in prediction of outcomes System cut-off Sensitivity Specificity AUC PLR NLR OR Major adverse cardiovascular event (MACE) GRACE ≤100 0.96 (0.90, 0.98) 0.26 (0.16, 0.40) 0.71 (0.67, 0.75) 1.3 (1.1, 1.5) 0.17 (0.10, 0.27) 8 (5, 12) >100 0.58 (0.53, 0.64) 0.69 (0.61, 0.77) 0.66 (0.62, 0.70) 1.9 (1.5, 2.3) 0.60 (0.54, 0.67) 3 (2, 4) HEART <4 0.99 (0.97, 0.99) 0.16 (0.08, 0.29) 0.98 (0.96, 0.99) 1.2 (1.0, 1.3) 0.09 (0.03, 0.25) 13 (4, 36) 4 0.93 (0.88, 0.96) 0.47 (0.40, 0.54) 0.72 (0.68, 0.76) 1.8 (1.6, 2.0) 0.14 (0.08, 0.23) 13 (8, 21) >4 0.77 (0.63, 0.87) 0.78 (0.59, 0.90) 0.84 (0.81, 0.87) 3.6 (1.8, 7.2) 0.29 (0.17, 0.49) 12 (5, 34) Acute myocardial infarction (AMI) GRACE NA 0.95 (0.86, 0.98) 0.29 (0.16, 0.46) 0.72 (0.68, 0.76) 1.3 (1.1, 1.6) 0.18 (0.10, 0.33) 7 (4, 12) HEART NA 0.94 (0.88, 0.97) 0.48 (0.32, 0.64) 0.86 (0.82, 0.88) 1.8 (1.4, 2.4) 0.12 (0.08, 0.21) 14 (9, 23) All-cause mortality GRACE NA 0.82 (0.75, 0.87) 0.51 (0.42, 0.61) 0.75 (0.71, 0.79) 1.7 (1.5, 1.9) 0.36 (0.29, 0.44) 5 (4, 6) HEART NA 0.78 (0.57, 0.90) 0.56 (0.49, 0.62) 0.65 (0.61, 0.69) 1.8 (1.3, 2.4) 0.39 (0.18, 0.89) 4 (1, 13) All measures are presented with 95% confidence interval. PLR: Positive likelihood ratio; NLR: Negative likelihood ratio; OR: Diagnostic odds ratio; AUC: Area under the curve; 95% CI: Confidence interval; NA: Not applicable. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 9 Archives of Academic Emergency Medicine. 2023; 11(1): e50 Figure 1: PRISMA flow diagram. Table 3: Risk of bias assessment Study, year Risk of Bias Applicability Overall Patient selection Index test Outcome Flow and timing Analysis Patient selection Index test Outcome Flow and timing Al-Zaiti, 2018 Low Low Low Unclear Unclear Low Low Low Low Some concern Carlton, 2015 Low Low Low Low Unclear Low Low Low Low Some concern Chae, 2016 Unclear Low Low Low Low Low Low Low Low Some concern Chen, 2016 Low Low Unclear Low Low Low Low Low Low Some concern Chew, 2018 Low Unclear Unclear Low Low Low Low Low Low Some concern Dinesh, 2022 Low Unclear Unclear Low Low Low Low Low Low Some concern Dupuy, 2021 Unclear Unclear Low Low Low Low Low Low Low Some concern Han, 2017 Low High Low Low Low Low Low Low Low Some concern Hrecko, 2022 Low Low Unclear Unclear Unclear Low Low Low Low Some concern Huang, 2021 Unclear Unclear Low Low Low Low Low Low Low Some concern Jukneviciene, 2022 Low Unclear Low Unclear Unclear Low Low Low Low Some concern Liu, 2017 High High Low Low Low Low Low Low Low Some concern Mingwei, 2020 High Low Low Low Low Low Low Low Low Some concern Poldervaar, 2016 Low Low Low Low Low Low Low Low Low Low Reaney, 2017 High Low Unclear Low Low Low Low Low Low Some concern Ruansomboon, 2020 Low Unclear Low Low Low Low Low Low Low Some concern Sakamoto, 2016 Unclear Low Unclear Low Low Low Low Low Low Some concern Shin, 2020 Low Low Unclear Low Low Low Low Low Low Some concern Singer, 2017 Unclear Unclear Low Low Unclear Low Low Low Low Some concern Steiro, 2021 Unclear Low Low Low Low Low Low Low Low Some concern Tekin, 2021 Unclear High High Low Low Low Low High Low Some concern Torralba, 2020 Unclear Low Low Low Low Low Low Low Low Some concern Wong, 2017 Low Unclear Low Low Low Low Low Low Low Some concern Yang, 2022 Low High Low Low Low Low Low Low Low Some concern Zheng, 2020 Low Low Low Low Low Low Low Low Low Some concern This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index A. Kabiri et al. 10 Figure 2: Performance of GRACE score with cut-off values of ≤ 100 (a) and > 100 (b) and HEART score with cut-off values < 4 (c), equal to 4 (d) and greater than 4 (e) in the prediction of major adverse cardiovascular events. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 11 Archives of Academic Emergency Medicine. 2023; 11(1): e50 Figure 3: Performance of GRACE score (a) and HEART score (b) in the prediction of acute myocardial infarction. Figure 4: Performance of GRACE score (a) and HEART score (b) in the prediction of all-cause mortality. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index A. Kabiri et al. 12 Figure 5: Publication bias of the included articles for the value of GRACE (a) and HEART (b) in the prediction of major adverse cardiovascular events. Table 4: Certainty of evidence Outcome Study/ population Study design Factors that may decrease certainty of evidence Certainty of evidence Risk of bias Indirectness Inconsistency Imprecision Publication bias Major adverse cardiovascular event GRACE Cut-offs ≤ 100 8 studies N = 10243 Cohort studies Serious Not serious Not serious Not serious Not serious ○ ⊕⊕⊕⊕ Moderate Cut-offs > 100 13 studies N = 10793 Cut-off < 4 2 studies N = 1431 HEART Cut-off = 4 14 studies N = 15239 Cohort studies Serious Not serious Not serious Not serious Not serious ○ ⊕⊕⊕⊕ Moderate Cut-offs > 4 6 studies N = 3814 Acute myocardial infarction GRACE Cut-offs: NA 6 studies N = 3591 Cohort studies Serious Not serious Not serious Not serious Not applicable ○ ⊕⊕⊕ ○ Low HEART Cut-off: NA 6 studies N = 3591 Cohort studies Serious Not serious Not serious Not serious Not applicable ○ ⊕⊕⊕ ○ Low All-cause mortality GRACE Cut-off: NA 3 studies N = 1903 Cohort studies Serious Not serious Not serious Not serious Not applicable ○ ⊕⊕⊕ ○ Low HEART Cut-off: NA 3 studies N = 1903 Cohort studies Serious Not serious Not serious Not serious Not applicable ○ ⊕⊕⊕ ○ Low This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index 13 Archives of Academic Emergency Medicine. 2023; 11(1): e50 Supplementary Table 1: Search strategy PubMed: ((Global registry acute coronary events[tiab] OR GRACE[tiab]) AND (HEART[tiab])) AND (SCOR*[tiab] OR Scal*[tiab] OR tool*[tiab] OR mode*[tiab] OR pathway[tiab]OR assessment*[tiab]) Embase: (’global registry of acute coronary events’/exp OR ’global registry of acute coronary events’:ab,ti OR ’grace’:ab,ti OR ’global registry for acute coronary events’:ab,ti) AND (’scor*’:ab,ti OR ’scal*’:ab,ti OR ’tool*’:ab,ti OR ’mode*’:ab,ti OR ’pathway’:ab,ti OR ’assess- ment*’:ab,ti) AND (’history electrocardiogram age risk factors and troponin score’/exp OR ’history electrocardiogram age risk factors and troponin score’ OR ’heart’:ab,ti) Scopus: TITLE-ABS-KEY(“global registry of acute coronary events” OR “grace” OR “global registry for acute coronary events”) AND TITLE-ABS- KEY( “scor*” OR “scal*” OR “tool*” OR “mode*” OR “pathway” OR “assessment*”) AND TITLE-ABS-KEY(“history electrocardiogram age risk factors and troponin score” OR “heart”) Web of Science (((TS=(“global registry of acute coronary events” OR “grace” OR “global registry for acute coronary events”))) AND TS=(“scor*” OR “scal*” OR “tool*” OR “mode*” OR “pathway” OR “assessment*”)) AND TS=(“history electrocardiogram age risk factors and troponin score” OR “heart”) This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: https://journals.sbmu.ac.ir/aaem/index.php/AAEM/index Introduction Methods Results Discussion Limitations Conclusion Declarations References