Emergency. 2017; 5 (1): e18 OR I G I N A L RE S E A RC H Screening Characteristics of TIMI Score in Predicting Acute Coronary Syndrome Outcome; a Diagnostic Accu- racy Study Mostafa Alavi-Moghaddam1, Saeed Safari2, Hamideh Alavi-Moghaddam1∗ 1. Emergency Department, Imam Hossein Hospital, Shahid Beheshti University of Medical Sciences, Tehran, Iran. 2. Emergency Department, Shohadaye Tajrish Hospital, Shahid Beheshti University of Medical Sciences, Tehran, Iran. Received: Jun 2016; Accepted: July 2016; Published online: 9 January 2017 Abstract: Introduction: In cases with potential diagnosis of ischemic chest pain, screening high risk patients for adverse outcomes would be very helpful. The present study was designed aiming to determine the diagnostic accuracy of thrombolysis in myocardial infarction (TIMI) score in Patients with potential diagnosis of ischemic chest pain. Methods: This diagnostic accuracy study was designed to evaluate the screening performance characteristics of TIMI score in predicting 30-day outcomes of mortality, myocardial infarction (MI), and need for revascular- ization in patients presenting to ED with complaint of typical chest pain and diagnosis of unstable angina or Non-ST elevation MI. Results: 901 patients with the mean age of 58.17 ± 15.00 years (19-90) were studied (52.9% male). Mean TIMI score of the studied patients was 0.97 ± 0.93 (0-5) and the highest frequency of the score belonged to 0 to 2 with 37.2%, 35.3%, and 21.4%, respectively. In total, 170 (18.8%) patients experienced the out- comes evaluated in this study. Total sensitivity, specificity, positive and negative predictive value, and positive and negative likelihood ratio of TIMI score were 20 (95% CI: 17 - 24), 99 (95% CI: 97 - 100), 98 (95% CI: 93 - 100), 42 (95% CI: 39 - 46), 58 (95% CI: 14 - 229), and 1.3 (95% CI: 1.2 - 1.4), respectively. Area under the ROC curve of this system for prediction of 30-day mortality, MI, and need for revascularization were 0.51 (95% CI: 0.47 - 0.55), 0.58 (95% CI: 0.54 - 0.62) and 0.56 (95% CI: 0.52 - 0.60), respectively. Conclusion: Based on the findings of the present study, it seems that TIMI score has a high specificity in predicting 30-day adverse outcomes of mortality, MI, and need for revascularization following acute coronary syndrome. However, since its sensitivity, negative predictive value, and negative likelihood ratio are low, it cannot be used as a proper screening tool for ruling out low risk patients in ED. Keywords: Coronary artery disease; prognosis; myocardial infarction; decision support techniques; angina, unstable © Copyright (2017) Shahid Beheshti University of Medical Sciences Cite this article as: Alavi-Moghaddam M, Safari S, Alavi-Moghaddam H, Kariman H. Screening Characteristics of TIMI Score in Predicting Acute Coronary Syndrome Outcome; a Diagnostic Accuracy Study. Emergency. 2017; 5 (1): e18. 1. Introduction Coronary artery disease (CAD) is a major health problem worldwide (1). Although during the past decades, with de- velopment of preventive measures and improvement of diag- nostic and therapeutic approaches, CAD related deaths have decreased by more than 40%, CAD still remains the largest killer of the US population (2, 3). About half of all deaths in developed countries and 25% of those in developing ones oc- ∗Corresponding Author: Hamideh Alavi-Moghaddam; Emergency Depart- ment, Imam Hossein Hospital, Shahid Madani Avenue, Imam Hossein Square, Tehran, Iran. Tel: 00989122030036, Email: hamide_mogaddam@yahoo.com cur due to CAD (4). It is predicted that until 2020, the num- ber of CAD related deaths will exceed that of infectious dis- eases (5). It seems that rapid diagnosis, and choosing the best treatment in the initial stages are of great importance in improving the outcome of CAD patients. In cases that elec- trocardiogram (ECG), as the most available and rapid diag- nostic tool, is not helpful in decision making (cases of un- stable angina and Non-ST elevation MI), screening high risk patients for adverse outcomes would be very helpful. Various clinical decision rules and scoring systems have been devel- oped for this purpose (6, 7). An ideal model in this regard, should have high sensitivity and be able to predict the out- come accurately, using readily available clinical information. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com M. Alavi-Moghaddam et al. 2 One of these systems is Thrombolysis In Myocardial Infarc- tion (TIMI) scoring model that helps divide the patients into 2 groups of low risk (score 0-1) and high risk (score > 1) re- garding adverse 30-day outcome. A study on 279 patients vis- iting ED showed that TIMI score is a valid tool for assessing 30-day mortality risk (8). Additionally, in a study on 3609 ST- elevation MI patients, it was concluded that TIMI score has both short and long term value in predicting patient mortal- ity in those that underwent primary percutaneous coronary intervention (9). A study by Lee et al. in 2011 in Pennsylvania also revealed the good ability of this system in predicting the incidence of 30-day cardiovascular adverse events following typical chest pain (10). Currently, in many developing coun- tries, supplementary diagnostic measures, such as exercise testing or echocardiography, is not available and cannot be done in EDs. In this situation, using clinical scoring systems may be helpful in patient disposition. Therefore, the present study was designed aiming to determine the diagnostic accu- racy of TIMI score in predicting 30-day outcome of patients presenting to ED with diagnosis of unstable angina or Non- ST elevation MI. 2. Methods 2.1. Study design This study was designed to evaluate the diagnostic accuracy of TIMI clinical scoring system in prediction of 30-day out- come of patients presenting to ED of Imam Hossein Hospi- tal, Tehran, Iran, with potential diagnosis of ischemic chest pain, during a 6-month period (from October 2015 to March 2016). The study was approved by the ethics committee of Shahid Beheshti University of Medical Sciences and the re- searchers adhered to the principles of Helsinki Declaration and patient data confidentiality throughout the study. A writ- ten informed consent was obtained from all the patients for participating in the study. 2.2. Participants All the patients with chest pain who were diagnosed with un- stable angina or Non-ST elevation MI were entered to the study without any age or sex limitation, using census sam- pling. Patients with unstable hemodynamics, evidence of ST segment elevation in the initial ECG and those who did not give consent for participating were excluded. Patients with diagnosis of ST elevation MI were immediately scheduled for receiving thrombolytic therapy or percutaneous coronary in- tervention (PCI). 2.3. Data collection After learning the study aims and method, the in charge emergency medicine residents were held responsible for data gathering and filling the checklist for the patients on admis- sion. All the patients were relocated to cardiac monitoring unit, and initial life support measures (pain relief with serum nitroglycerin and intravenous morphine, O2 therapy, blood pressure monitoring and . . . ) were initiated on admission to ED. In addition, ECG and cardiac enzymes were ordered. All patients underwent 12 lead ECG and in cases suspected to inferior or posterior wall MI, posterior/inferior leads were added. If the pain was not relieved within 20 - 30 minutes after admission, despite initiating therapy with nitroglycerin and intravenous morphine, ECG was repeated with 20 -30 minute intervals up to 3 times. ECG interpretations were done by emergency medicine residents and approved by the in charge attend of the shift and on call internists. Based on the treatment protocol of the studied health center, eval- uation of cardiac enzymes including creatine kinase (CK), CKmb, and troponin was also ordered for all the patients on admission and 6 hours later. If the final diagnosis was un- stable angina or non-ST elevation MI, all the baseline char- acteristics of the patient including age and sex, pain char- acteristics, status of the known risk factors of arteriosclero- sis (hypertension, diabetes, smoking, overweight, hyperlipi- demia and . . . ), history of cardiovascular diseases (stroke, MI, and . . . ), drugs history, ECG findings, laboratory findings es- pecially level of cardiac enzymes as well as variables needed for calculating TIMI score were gathered by the in charge res- ident, using a checklist prepared for this purpose. Required data regarding the final outcome of the patients were gath- ered from their clinical files or, if not available, by phone call with the patient, their relatives, or their cardiologist by a senior emergency medicine resident. Considering the def- inition of TIMI score the evaluated outcomes including MI, need for revascularization and all-cause mortality during the 30 days after admission to ED (11). 2.4. Statistical Analysis All analyses were done using SPSS version 20 and STATA 11. Mean, standard deviation, median, range, frequency, and percentage were used to describe data. To calculate the diag- nostic accuracy of TIMI system, sensitivity, specificity, posi- tive and negative predictive value, and positive and negative likelihood ratio, and area under the receiver operating char- acteristic (ROC) curve were calculated. 3. Results: 3.1. Baseline characteristics of the patients 901 patients with the mean age of 58.17 ± 15.00 years (19- 90) were studied (52.9% male). Table 1 shows the baseline characteristic of the patients. The most common accompa- nying diseases included hypertension (51.4%) and diabetes (19.8%). ST segment changes (39.2%) and T wave inver- sion (31.6%) were the most common early ECG findings, re- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com 3 Emergency. 2017; 5 (1): e18 Table 1: Baseline characteristics of the studied patients Variable Number (%) Sex Female 424 (47.1) Male 477 (42.9) Risk factors Hypertension (51.4) 463 Diabetes 178 (19.8) Smoking 88 (9.8) Positive family history 29 (3.2) Hyperlipidemia 88 (9.8) Number of accompanying risk factors 0 289 (32.1) 1 414 (45.9) 2 164 (18.2) 3 32 (3.6) 4 2 (0.2) History of cardiac failure 26 (2.9) PTCA history 58 (6.4) ECG findings ST deviation 353 (39.2) T inversion 285 (31.6) Left bundle branch block 28 (3.1) Right bundle branch block 3 (0.3) Atrial fibrillation 17 (1.2) Sinus tachycardia 8 (0.9) Ventricular tachycardia 1 (0.1) Multifocal atrial tachycardia 1 (0.1) Paroxysmal supraventricular tachycardia 6 (0.7) PTCA: Percutaneous transluminal coronary angioplasty; ECG: electrocardiography. Table 2: Frequency of TIMI risk factors among studied patients (n = 901) Item Number (%) Age ≥ 65 321 (35.6) ≥ 3 CAD risk factors 34 (3.8) Known CAD (stenosis ≥ 50%) 12 (1.3) ASA use in past 7 days 59 (6.5) Sever angina (≥ 2 episodes / 24 hours) 3 (0.3) ST segment changes ≥ 0.5 mm 353 (39.2) Positive cardiac marker 95 (10.5) CAD: Coronary artery disease; ASA: Acetylsalicylic acid. Table 3: 30-day outcomes of studied patients Outcome Number (%) Myocardial infarction 88 (9.8) Urgent revascularization 75 (8.3) Coronary angiography 488 (54.2) Percutaneous coronary intervention (PCI) 332 (36.8) Coronary artery bypass graft (CABG) 34 (3.8) Admission to coronary care unit (CCU) 712 (79.0) Death 7 (0.8) spectively. Table 2 depicts the frequency of factors used in TIMI clinical scoring system among the studied patients. The most common factors in this regard were more than 0.5 mm changes in ST segment (39.2%) and age ≥ 65 years (35.6%). Mean TIMI score of the studied patients was 0.97 ± 0.93 (0-5) and the highest frequency of the score belonged to 0 to 2 with 37.2%, 35.3%, and 21.4%, respectively. Table 3 reveals the fi- nal outcome of the participants. 488 (54.2%) of the patients underwent angiography of coronary arteries, based on which 22 (2.4%) had severe obstruction, 371 (41.2%) had moder- ate obstruction, 72 (8.0%) had mild and 23 (2.6%) had non- significant obstruction. 3.2. Characteristics of TIMI screening system In total, 170 (18.8%) patients experienced the outcomes evaluated in this study. Total sensitivity, specificity, positive and negative predictive value, and positive and negative likelihood ratio of TIMI score were 20 (95% CI: 17 - 24), 99 (95% CI: 97 - 100), 98 (95% CI: 93 - 100), 42 (95% CI: 39 - 46), 58 (95% CI: 14 - 229), and 1.3 (95% CI: 1.2 - 1.4), respectively. Table 4 shows the screening characteristics of TIMI score for predicting 30-day mortality, need for revascularization, and MI. The findings reveal the high specificity and low sensitivity of this system in these regards. Area under the ROC curve of this system for prediction of 30-day mortality, MI, and need for revascularization were 0.51 (95% CI: 0.47 - 0.55), 0.58 (95% CI: 0.54 - 0.62) and 0.56 (95% CI: 0.52 - 0.60), respectively. 4. Discussion: Findings of the present study, reveal the high specificity (99%) of TIMI score in prediction of 30-day clinical outcomes of mortality, MI, and need for revascularization. However, it has a low sensitivity (1.2 - 15%) and its screening power is low. A study by Marcoon et al. in 2013 in America showed that the probability of adverse 30-day outcome in patients with typi- cal chest pain and 0 TIMI score is lower than 1% (12). In addi- tion, a cohort study by Pollack et al. revealed that the risk of short term adverse outcome is 2.1% in patients with 0 TIMI score and 100% in those with a score higher than 7 (13). In a meta-analysis, Hess et al. expressed that there is 1.8% risk of 30-day cardiac events in those with 0 TIMI score, while the risk is 4% for those with a score of 1. They found a strong linear correlation between TIMI score and incidence of car- diac events (14). In a cohort study in 30, Lee et al. concluded that patients with high risk of adverse outcome may be iden- tified in ED using this system (10). This result is completely in line with the findings of the present study since the high specificity of this model makes it a useful tool to rule in at risk patients. On the other hand, low sensitivity, negative pre- This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com M. Alavi-Moghaddam et al. 4 Table 4: Screening performance characteristics of TIMI score in 30-day outcome prediction of patients with unstable angina/non-ST eleva- tion myocardial infarction (MI) Outcomes Mortality MI Revascularization Sensitivity 1.2 (0.5-2.6) 15 (12-18) 13 (10-16) Specificity 100 (98-100) 100 (98-100) 99 (97-99) Positive predictive value 100 (56-100) 100 (94-100) 97 (89-99) Negative predictive value 38 (35-42) 41 (37-44) 40 (36-43) Positive likelihood ratio ∞ ∞ 37 (9-143) Negative likelihood ratio 1.5 (1.4-1.6) 1.4 (1.3-1.5) 1.5 (1.4-1.6) Data were presented with 95% confidence interval. dictive value and negative likelihood ratio has limited its use for screening and ruling out high risk patients in ED. How- ever, this tool can be at least used for preventing discharge of high risk patients from ED. In the meta-analysis by Hess et al. that included 8 studies, at >0 cut off, sensitivity was cal- culated to be 97.2%, specificity was 25.0%, and negative like- lihood ratio was 0.11. They concluded that patients should not be discharged solely because of low TIMI score (14). The findings of the preset study also showed the low sensitivity (20%) and high specificity (99%) of TIMI score. Jain et al. compared the prognostic value of TIMI score with HEART score in predicting 30-day and 5-year outcomes of mortality, MI, and revascularization, and concluded that HEART score is superior (19). This finding was in line with other studies that expressed HEART score has higher discriminatory power compared to TIMI score (20, 21). It has also been found that Global Registry of Acute Coronary Events (GRACE) risk score has significantly higher discriminatory accuracy in compari- son with TIMI score (22). In comparison of HEART, TIMI, and GRACE scores, HEART score is found to be the best of the 3 (23). A study compared Emergency Department Assessment of Chest Pain Score (EDACS), with HEART and TIMI scores in prediction of 6-week major adverse cardiovascular events of MI, revascularization and death, in a 6 week follow-up pe- riod for patients with chest pain presenting to ED. It was re- vealed that EDACS score was able to identify more low-risk patients compared to the other 2 scores (24). Modified Gold- man risk score in combination with high-sensitivity troponin was also found to identify significantly more low-risk patients for 30-day adverse outcomes compared to TIMI score (25). CHADS2 score has also been deemed more practical for out- come prediction of acute MI patients (26). Some studies even expressed that TIMI score does not have a predictive capac- ity, since its area under the ROC curve is 0.532, which is close to the area under the curve found in this study (27). Among the limitations of this study is extraction of outcome from pa- tients’ medical profile might leading to miss some important points, however, we tried to minimize this probability by con- tacting the patient or their in-charge cardiologist. 5. Conclusion: Based on the findings of the present study, it seems that TIMI score has a high specificity in predicting 30-day adverse out- comes of mortality, MI, and need for revascularization fol- lowing acute coronary syndrome. However, since its sensi- tivity, negative predictive value, and negative likelihood ratio are low, it cannot be used as a proper screening tool for ruling out low risk patients in ED. 6. Appendix 6.1. Acknowledgements This article has been derived from Dr. Hamideh Alavi- Moghaddam student thesis to receive his specialist degree in Emergency Medicine from Shahid Beheshti University of Medical Sciences. Authors would like to acknowledge all the emergency department staff of Imam Hossein and Shoha- daye Tajrish Hospitals. 6.2. Authors contribution All authors made a substantial contribution to analysis and writing of the paper draft and met the 4 criteria of author- ship recommended by the International Committee of Med- ical Journal Editors. 6.3. Conflict of interest None. 6.4. Funding None. References 1. Hata J, Kiyohara Y. Epidemiology of stroke and coronary artery disease in Asia. Circulation Journal. 2013;77(8):1923-32. 2. Thom T, Haase N, Rosamond W, et al. Heart disease and stroke statistics–2006 update: a report from the Amer- ican Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation. 2006;113(6):e85. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com 5 Emergency. 2017; 5 (1): e18 3. Ford ES, Ajani UA, Croft JB, et al. Explaining the de- crease in US deaths from coronary disease, 1980-2000. New England Journal of Medicine. 2007;356(23):2388-98. 4. Beaglehole R. International trends in coronary heart dis- ease mortality and incidence rates. European Journal of Cardiovascular Risk. 1999;6(2):63-8. 5. Huffman MD, Lloyd-Jones DM, Ning H, et al. Quan- tifying Options for Reducing Coronary Heart Dis- ease Mortality By 2020Clinical Perspective. Circulation. 2013;127(25):2477-84. 6. Montalescot G, Sechtem U, Achenbach S, et al. 2013 ESC guidelines on the management of stable coronary artery disease. European heart journal. 2013;34(38):2949-3003. 7. Roffi M, Patrono C, Collet J-P, et al. 2015 ESC Guidelines for the management of acute coronary syndromes in pa- tients presenting without persistent ST-segment eleva- tion. European heart journal. 2015:ehv320. 8. Ehsan MA, Mahmood M, Siddique MA, et al. Prediction of Major Adverse Cardiac Events of Patients with Acute Coronary Syndrome by Using TIMI Risk Score. University Heart Journal. 2013;8(2):73-9. 9. Damman P, Woudstra P, Kuijt WJ, et al. ShortâĂŘand LongâĂŘTerm Prognostic Value of the TIMI Risk Score after Primary Percutaneous Coronary Intervention for STâĂŘsegment Elevation Myocardial Infarction. Journal of interventional cardiology. 2013;26(1):8-13. 10. Lee B, Chang AM, Matsuura AC, Marcoon S, Hollander JE. Comparison of cardiac risk scores in ED patients with potential acute coronary syndrome. Critical pathways in cardiology. 2011;10(2):64-8. 11. Antman EM, Cohen M, Bernink PJ, 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. 12. Marcoon S, Chang AM, Lee B, Salhi R, Hollander JE. HEART score to further risk stratify patients with low TIMI scores. Critical pathways in cardiology. 2013;12(1):1-5. 13. Pollack CV, Sites FD, Shofer FS, Sease KL, Hollander JE. Application of the TIMI risk score for unstable angina and nonâĂŘST elevation acute coronary syndrome to an unselected emergency department chest pain popula- tion. Academic emergency medicine. 2006;13(1):13-8. 14. Hess EP, Agarwal D, Chandra S, et al. Diagnostic accuracy of the TIMI risk score in patients with chest pain in the emergency department: a meta-analysis. Canadian Med- ical Association Journal. 2010;182(10):1039-44. This open-access article distributed under the terms of the Creative Commons Attribution NonCommercial 3.0 License (CC BY-NC 3.0). Downloaded from: www.jemerg.com Introduction Methods Results: Discussion: Conclusion: Appendix References