A Retrospective Validation of the HEAR Score Authors: Jordan Luli, Andrew Scheidemantel, Katelyn Bennett, Jessalyn Schafer, Andrew Costa, Marie Lockhart, Kevin Stiver Published: 2024 ABSTRACT Studies have proposed the history, electrocardiogram, age, risk factors (HEAR) score to risk-stratify patients with chest pain in the emergency department. This paper assessed the utility of the HEAR score for ambulatory patients with chest pain. The findings demonstrate patients with low and high HEAR scores can be safely managed in the outpatient setting. Chest pain is one of the most common reasons for patients to present to the emergency department (ED). The American College of Cardiology/American Heart Association (ACC/AHA) 2021 chest pain guidelines1 outline clinical decision pathways to risk-stratify patients presenting with chest pain to the ED. The history, ECG, age, risk factors, troponin (HEART) score has been supported by these guidelines to risk-stratify patients with chest pain in the United States. Data have shown that patients with a HEART score ≤3 are at very low risk for an adverse event at 30 days and can be monitored in the outpatient setting.2Take-Home Message•Patients with low and high HEAR scores can be safely managed in an outpatient/ambulatory setting, which is more cost effective while maintaining patient safety. Take-Home Message •Patients with low and high HEAR scores can be safely managed in an outpatient/ambulatory setting, which is more cost effective while maintaining patient safety. •Patients with low and high HEAR scores can be safely managed in an outpatient/ambulatory setting, which is more cost effective while maintaining patient safety. Patients with low and high HEAR scores can be safely managed in an outpatient/ambulatory setting, which is more cost effective while maintaining patient safety. Recent studies have proposed the history, ECG, age, risk factors (HEAR) score as an alternative tool to risk-stratify patients with very low risk chest pain in the ED. In 2 studies, patients in the ED with HEAR scores <2 were considered safe for ambulatory evaluation.3,4 Patients with chest pain are often referred for further ischemic testing in the ambulatory setting, but little evidence exists for a risk-stratifying tool to evaluate appropriateness of this strategy. Troponins are rarely obtained in the ambulatory setting; therefore, the HEAR score may be an attractive tool for clinicians. We assessed the utility of the HEAR score for patients presenting with chest pain in the ambulatory setting and evaluated their outcomes. Objectives To assess the reliability of the HEAR score as an alternative tool to risk-stratify patients with chest pain using retrospective data from ambulatory cardiology office visits. Methods Our study was a single health care system retrospective chart review performed for new patient consultations who presented to cardiology offices from January 1, 2018, through December 31, 2021. The study was reviewed and deemed exempt by our institution’s institutional review board. We identified potential patients by using ICD-10 codes for chest pain linked to referral orders. Study data were collected using REDCap data capture tools hosted at the OhioHealth Research Institute.5,6 The demographics of each patient at the time of the cardiology visit were collected (Table 1). The components of the HEAR score were recorded by the evaluator using visit documentation. The most recent electrocardiogram (ECG) prior to the cardiology visit was used in the score if an ECG was not obtained during that visit.Table 1Demographics and Medical History for Patient Population (N = 278)Age, y57 (45.3-66.8)Sex Female151 (54.3) Male127 (45.7)Race African American/Black9 (3.2) Caucasian261 (94) Othera8 (2.9)Ethnicity Declined3 (1.1) Hispanic or Latino5 (1.8) Not Hispanic or Latino270 (97.1)Coronary artery disease62 (22.3)Previous MI29 (10.4)Previous PCI or CABG48 (17.3)Atherosclerotic disease54 (19.4)Taking aspirin90 (32.4)Peripheral artery disease8 (2.9)HEAR score specified risk factors Diabetes61 (21.9) Smoking history Current76 (27.3) Former78 (28.1) Never124 (44.6) Family history of CAD39 (14) Hypercholesteremia143 (51.4) Hypertension173 (62.2) BMI >30 kg/m2161 (57.9)Values are median Q1-Q3 or n (%).CABG = coronary artery bypass grafting; CAD = coronary artery disease; MI = myocardial infarction; PCI = percutaneous coronary intervention.aOther is a combined category of Asian, American Indian or Alaskan Native, declined to specify, and 2 or more races. Table 1Demographics and Medical History for Patient Population (N = 278)Age, y57 (45.3-66.8)Sex Female151 (54.3) Male127 (45.7)Race African American/Black9 (3.2) Caucasian261 (94) Othera8 (2.9)Ethnicity Declined3 (1.1) Hispanic or Latino5 (1.8) Not Hispanic or Latino270 (97.1)Coronary artery disease62 (22.3)Previous MI29 (10.4)Previous PCI or CABG48 (17.3)Atherosclerotic disease54 (19.4)Taking aspirin90 (32.4)Peripheral artery disease8 (2.9)HEAR score specified risk factors Diabetes61 (21.9) Smoking history Current76 (27.3) Former78 (28.1) Never124 (44.6) Family history of CAD39 (14) Hypercholesteremia143 (51.4) Hypertension173 (62.2) BMI >30 kg/m2161 (57.9)Values are median Q1-Q3 or n (%).CABG = coronary artery bypass grafting; CAD = coronary artery disease; MI = myocardial infarction; PCI = percutaneous coronary intervention.aOther is a combined category of Asian, American Indian or Alaskan Native, declined to specify, and 2 or more races. Demographics and Medical History for Patient Population (N = 278) Values are median Q1-Q3 or n (%). CABG = coronary artery bypass grafting; CAD = coronary artery disease; MI = myocardial infarction; PCI = percutaneous coronary intervention. Other is a combined category of Asian, American Indian or Alaskan Native, declined to specify, and 2 or more races. Further ischemic testing was recorded based on the recommendations of the cardiologist at the end of the visit. Major adverse cardiac events (MACEs) were recorded within 6 weeks of the initial evaluation and were defined as follows: acute myocardial infarction, unstable angina, stroke, left heart catheterizations (LHCs) with or without stenting, coronary artery bypass grafting (CABG) surgery, and/or cardiovascular related death. We tracked ED visits within our health care system related to chest pain over the same period. Continuous data were reported as medians with interquartile ranges due to a lack of normal distribution. Shapiro-Wilk, chi-square, and Mann-Whitney U tests were used. Statistical significance was set at P < 0.05. All comparison tests were performed using JASP software version 0.17.1 (JASP Team) and R version 4.2.2 (R Core Team). MedCalc was used to calculate positive and negative predictive values with a presumed prevalence of ischemic disease in the outpatient population at 10%.7 Results Out of 349 charts, 278 met inclusion criteria and were reviewed for HEAR scoring. Those excluded from the study did not have matching ICD-10 codes for a cardiac diagnosis or were younger than 18 years of age at the time of referral. The evaluating cardiologist billed for inclusion criteria ICD-10 codes at 96.4% of the visits. The median age of the patient population was 57 years and included more females (54.3%) than males (45.7%). Most patients were White (94%). Fifty-four percent of all patients had active or previous tobacco use, 51.4% had hypercholesterolemia, 62.2% had hypertension, and 57.9% had an elevated BMI (>30 kg/m2). Only 21.9% of patients had diabetes. The median HEAR score was 3.4 (for HEAR score distribution, see Table 2). A total of 218 patients were referred for ischemic evaluation; 103 patients had a HEAR score ≤3 vs 115 patients with a HEAR score ≥4 (P = 0.027). Of the patients, 174 were referred for stress testing, 27 were referred for outpatient LHC, and 13 were referred for outpatient coronary computed tomography angiography. Twenty nine patients were referred for further testing but did not complete the requested testing or the documentation was unavailable. The median time between cardiologist evaluation and completion of the testing was 14 days. All ischemic testing was obtained in the ambulatory setting.Table 2HEAR Score Distribution for Patient Population (N = 278)08 (2.9)131 (11.2)244 (15.8)358 (20.9)453 (19.1)561 (21.9)618 (6.5)75 (1.8)80Values are n (%). Median HEAR score of 3.4. Table 2HEAR Score Distribution for Patient Population (N = 278)08 (2.9)131 (11.2)244 (15.8)358 (20.9)453 (19.1)561 (21.9)618 (6.5)75 (1.8)80Values are n (%). Median HEAR score of 3.4. HEAR Score Distribution for Patient Population (N = 278) Values are n (%). Median HEAR score of 3.4. Twenty patients had an MACE within 6 weeks of evaluation. Three patients with HEAR scores of 0 to 3 experienced an MACE (2 patients had an LHC with one stented and another diagnosed with unstable angina). For patients with HEAR scores ≥4, 17 patients experienced an MACE (1 experienced acute myocardial infarction, 2 were diagnosed with unstable angina, 5 underwent CABG, and 13 had LHC with 7 having stents placed) (P < 0.001). Most MACE events were LHC with 2 patients with HEAR scores ≤3 and 13 patients with HEAR scores ≥4 (P = 0.003). A total of 11 patients were evaluated in the ED with a complaint of chest pain within 6 weeks of cardiology evaluation: 4 patients with HEAR score ≤3 and 7 with HEAR score ≥4. Of these, 8 were determined to be noncardiac, 1 was evaluated with an inpatient stress test only, and 2 underwent LHC with one of these patients receiving a stent. These latter 3 patients accounted for each of the 3 patients with acute coronary syndrome within our MACE counts. When comparing the HEAR score in low-risk patients (HEAR score ≤3) with intermediate- to high-risk patients (HEAR score ≥4), the sensitivity and specificity of identifying ischemic disease was 73.3% (95% CI: 58.1%-85.4%) and 55.8% (95% CI: 49.2%-62.4%), respectively. The presence of ischemic disease was defined as intermediate- or high-risk stress test results, presence of disease on coronary computed tomography angiography or LHC, and/or MACE within 6 weeks of cardiology visit. Assuming a 10% prevalence of ischemic disease in an outpatient cardiology clinic, the positive predictive value and negative predictive values were 15.6% (95% CI: 9.9%-22.8%) and 95% (95% CI: 90%-98%) respectively (Table 3).Table 3Cardiology Visit Characteristics for Patient Population (N = 278)Cardiologist billed for the referral diagnosis code260 (96.4)Most recent ECG interpretation Non-specific changes79 (28.4) Normal193 (69.4) ST-segment depressionsa6 (2.2)History score documentation Cardiologist83 (29.9) Chart reviewer195 (70.1)Patient referred for further ischemic evaluation218 (78.4)Ischemic evaluation performed, n = 218 Referred but no documentation29 (13.3) Stress test174 (80) CCTA13 (6) LHC27 (12.4)Days between cardiology visit and ischemic evaluation14 (9.5-22)Vitals Height (cm)168.9 (160-177.8) Weight (kg)89.2 (75.8-107.7) BMI (kg/m²)32.4 (27.3-36.3) Systolic BP (mm Hg)130 (120-143) Diastolic BP (mm Hg)81 (76-86) Heart Rate (beats/min)75 (67-84)Values are n (%) or median (Q1-Q3).CCTA = coronary computed tomography angiography; ECG = electrocardiogram; LHC = left heart catheterization.aST-segment depressions not due to LVH, LVBB, or digox-in use. Table 3Cardiology Visit Characteristics for Patient Population (N = 278)Cardiologist billed for the referral diagnosis code260 (96.4)Most recent ECG interpretation Non-specific changes79 (28.4) Normal193 (69.4) ST-segment depressionsa6 (2.2)History score documentation Cardiologist83 (29.9) Chart reviewer195 (70.1)Patient referred for further ischemic evaluation218 (78.4)Ischemic evaluation performed, n = 218 Referred but no documentation29 (13.3) Stress test174 (80) CCTA13 (6) LHC27 (12.4)Days between cardiology visit and ischemic evaluation14 (9.5-22)Vitals Height (cm)168.9 (160-177.8) Weight (kg)89.2 (75.8-107.7) BMI (kg/m²)32.4 (27.3-36.3) Systolic BP (mm Hg)130 (120-143) Diastolic BP (mm Hg)81 (76-86) Heart Rate (beats/min)75 (67-84)Values are n (%) or median (Q1-Q3).CCTA = coronary computed tomography angiography; ECG = electrocardiogram; LHC = left heart catheterization.aST-segment depressions not due to LVH, LVBB, or digox-in use. Cardiology Visit Characteristics for Patient Population (N = 278) Values are n (%) or median (Q1-Q3). CCTA = coronary computed tomography angiography; ECG = electrocardiogram; LHC = left heart catheterization. ST-segment depressions not due to LVH, LVBB, or digox-in use. Discussion Guidelines have been developed to provide clinical decision pathways on how to risk-stratify patients with chest pain who present to the ED, but there are little data on how to manage these patients in the ambulatory setting. Furthermore, there is a lack of evidence when comparing risk profiles of patients with chest pain who present to the ED and cardiology offices. We found patients with a low HEAR score (≤3) to receive ischemic testing at a statistically significant lower rate than patients with an intermediate or high HEAR score. This is consistent with ACC/AHA 2021 guidelines. Along with the lower ischemic testing rate, the MACE rates were significantly lower in the low-risk patients. Given the lower ischemic testing rate, lower MACE rates, and the high negative predictive value, a low HEAR score in the ambulatory setting could be used to rule out ischemic disease and avoid unnecessary testing. For patients with an elevated HEAR score (HEAR score ≥4), we found no significant difference in ED visits and cardiovascular or noncardiovascular deaths. These patients did have a higher rate of MACE as seen by 17 patients vs only 3 with low HEAR score (≤3) (Table 4). We are reassured that a numerical increase in acute coronary syndrome was not seen because only 1 patient in each group presented with acute myocardial infarction and there was no significant difference in patients with unstable angina at 6 weeks. These patients would likely have been admitted to the hospital per 2021 ACC/AHA guidelines, in which admission is recommended in the intermediate- and high risk-populations. We found that these higher-risk patients were safely monitored in the outpatient setting.Table 4Outcomes at 6 weeks for HEAR Score ≤3 vs ≥4HEAR ≤3(n = 141)HEAR ≥24(n = 137)P ValueCardio-related ED visit4 (2.8)7 (5.1)0.331All MACEa3 (2.1)17 (12.6)<0.001MACE Categories Acute MI1 (0.7)1 (0.7)0.975 Unstable angina1 (0.7)2 (1.5)0.383 Stroke0 (0)0 (0)– LHCb2 (1.4)13 (9.6)0.003 CABG0 (0)5 (3.7)0.021 Cardiovascular-related death0 (0)0 (0)–Noncardiovascular-related death0 (0)1 (0.7)0.723Patient referred for ischemic evaluation103 (73.1)115 (83.9)0.027Stress test result, n = 169 High risk2 (2.4)7 (8.3) Intermediate risk5 (5.9)11 (13.1)0.194 Low risk24 (28.2)20 (23.8) Non diagnostic3 (3.5)3 (3.6) Normal51 (60)43 (51.2)Presence of disease on CCTA or LHC, n = 577/37 (35)26/20 (70.3)0.01Values are n (%).CCTA = coronary computed tomography angiography; ED = emergency department; LHC = left heart catheterization; MACE = major adverse cardiovascular events.aMACE is a composite of Acute MI, unstable angina, stroke, LHC with or without stent, CABG, and cardiovascular-related death.bLHC with or without stent. Table 4Outcomes at 6 weeks for HEAR Score ≤3 vs ≥4HEAR ≤3(n = 141)HEAR ≥24(n = 137)P ValueCardio-related ED visit4 (2.8)7 (5.1)0.331All MACEa3 (2.1)17 (12.6)<0.001MACE Categories Acute MI1 (0.7)1 (0.7)0.975 Unstable angina1 (0.7)2 (1.5)0.383 Stroke0 (0)0 (0)– LHCb2 (1.4)13 (9.6)0.003 CABG0 (0)5 (3.7)0.021 Cardiovascular-related death0 (0)0 (0)–Noncardiovascular-related death0 (0)1 (0.7)0.723Patient referred for ischemic evaluation103 (73.1)115 (83.9)0.027Stress test result, n = 169 High risk2 (2.4)7 (8.3) Intermediate risk5 (5.9)11 (13.1)0.194 Low risk24 (28.2)20 (23.8) Non diagnostic3 (3.5)3 (3.6) Normal51 (60)43 (51.2)Presence of disease on CCTA or LHC, n = 577/37 (35)26/20 (70.3)0.01Values are n (%).CCTA = coronary computed tomography angiography; ED = emergency department; LHC = left heart catheterization; MACE = major adverse cardiovascular events.aMACE is a composite of Acute MI, unstable angina, stroke, LHC with or without stent, CABG, and cardiovascular-related death.bLHC with or without stent. Outcomes at 6 weeks for HEAR Score ≤3 vs ≥4 Values are n (%). CCTA = coronary computed tomography angiography; ED = emergency department; LHC = left heart catheterization; MACE = major adverse cardiovascular events. MACE is a composite of Acute MI, unstable angina, stroke, LHC with or without stent, CABG, and cardiovascular-related death. LHC with or without stent. Another reassuring finding is the low rate of percutaneous coronary intervention in those undergoing LHC within 6 weeks of the cardiology visit. Two patients with low HEAR scores underwent LHC with only 1 requiring stenting. In the high HEAR score group, 7 of 13 patients underwent stenting and 2 were referred for CABG. This suggests that patients with elevated HEAR scores can be safely managed in the outpatient setting when it comes to LHC despite a likely admission should they have presented through the ED. This was a single-center, retrospective study with potential limitations due to interobserver variability and only tracking ED visits within our health care system. We attempted to limit the interobserver variability by reviewing the first 10 patients as a group with a board-certified cardiologist. The data on LHC may be artificially elevated if the reviewer considered this to be a MACE event even though it was a result of the initial ischemic evaluation. If troponins, especially high-sensitivity troponins, were drawn at the time of the visit, this may have changed management by elevating the HEART score. Conclusions The HEAR score is a simple and fast tool that is routinely used in EDs and supported by guidelines to triage patients with chest pain. Our health care system saw that the HEAR score is a safe tool for managing ambulatory chest pain patients, but more retrospective data need to be collected. This could have greater implications on how we triage patients with elevated risk scores in an ambulatory setting that is more cost effective while maintaining patient safety. Funding Support and Author Disclosures The authors have reported that they have no relationships relevant to the contents of this paper to disclose.