Clinical background
Chest pain is one of the commonest reasons for attending an emergency department, and management turns on separating patients with acute coronary syndrome from those with non-cardiac causes. Troponin testing is central to this triage, but as many as half of all troponin requests in emergency departments are judged unnecessary, which prolongs the stay, drives costs and can lead to overdiagnosis without any gain in safety [3]. The HEAR score was developed to identify a group at such low risk that deferring or omitting troponin testing altogether can be considered, and thereby to speed up management for patients who can probably be discharged directly.
The score is derived from the HEART score by removing the troponin component. Its purpose is not to replace full risk stratification, but to act as a preliminary step determining whether biomarker testing is needed at all.
Calculating the HEAR score
The HEAR score comprises four components, each scored 0 to 2, giving a total range from 0 to 8:
where:
- H (History): Slightly suspicious = 0, Moderately suspicious = 1, Highly suspicious = 2. The assessment of the history concerns how well the symptom pattern matches typical infarction pain, taking account of character, site, radiation and provocation.
- E (ECG): Normal = 0, Non-specific repolarisation disturbance = 1, Significant ST-segment deviation = 2.
- A (Age): <45 years = 0, 45–64 years = 1, ≥65 years = 2.
- R (Risk factors): No known risk factors = 0, 1–2 risk factors = 1, ≥3 risk factors or known atherosclerotic disease = 2. The risk factors are hypertension, hypercholesterolaemia, diabetes, obesity (BMI >30), smoking (current or within the last 3 months), family history, or known atherosclerotic disease.
The HEAR score was not derived as a standalone model development, but by removing the troponin component from the established HEART score, which was originally developed in a Dutch emergency department cohort. The first large validation of the HEAR score as an independent instrument was carried out by Moumneh et al. in a retrospective cohort at 15 emergency departments within Kaiser Permanente Southern California between May 2016 and December 2017 [1]. The study comprised 22,109 patient visits in which the treating physician had documented a HEART score, and the HEAR score was calculated retrospectively by excluding the troponin value. The outcome measure was acute myocardial infarction or all-cause death within 30 days. Patients with ST-elevation infarction, hospice care or "do not resuscitate" status were excluded.
Interpretation in practice
The only clinically decisive threshold for the HEAR score is ≤1 point, which identifies a group at very low risk. A score ≥2 does not mean high risk; it marks only that troponin testing, and hence the full HEART score, is still needed.
| HEAR score | Clinical interpretation | Action |
|---|---|---|
| 0–1 | Very low risk of 30-day myocardial infarction/death | Consider omitting troponin testing. The patient can potentially be discharged from the emergency department without further cardiac investigation, provided clinical assessment and the history support that conclusion. |
| ≥2 | Not in itself high risk | Troponin testing indicated. A full HEART score or other risk stratification should be carried out before deciding on discharge or admission. |
In the Kaiser Permanente validation cohort, patients with a HEAR score ≤1 made up approximately 19% of all visits (4,106 of 22,109), and within this group 30-day myocardial infarction or death occurred in 0.1% (5 of 4,106; 95% CI 0.1–0.3) [1]. In the multisite validation carried out by Ashburn et al. at five emergency departments in North Carolina, patients with a HEAR score ≤1 made up 17.2% of the cohort (1,565 of 9,105), and 30-day MACE (death, myocardial infarction or revascularisation) occurred in 0.7% (11 of 1,565; 95% CI 0.4–1.3) [2]. The negative predictive value for 30-day MACE was 99.3% (95% CI 98.7–99.6) [2].
Validation and performance
The primary validation of the HEAR score was thus carried out by Moumneh et al. in the Kaiser Permanente cohort (22,109 patients), in which sensitivity for 30-day myocardial infarction or death was 97.9% and specificity 18.8% at the threshold of ≤1 [1]. The low specificity is expected and is not a sign of poor performance: the instrument is designed to identify low-risk patients, not to screen out high-risk patients.
Ashburn et al. carried out a prospective multisite validation at five emergency departments in North Carolina between November 2020 and July 2022 [2]. The study included 9,105 patients without known coronary artery disease and with a non-ischaemic ECG, in whom the treating physician calculated the HEAR score prospectively. Sensitivity for 30-day MACE at a HEAR score ≤1 was 97.9% (95% CI 96.2–98.9) and the NPV 99.3% (95% CI 98.7–99.6) [2]. When high-sensitivity troponin was added to the strategy, the NPV rose to 99.8%, but the net reclassification improvement was only 0.7% (95% CI −0.4 to 1.8; p = 0.21), that is, the troponin did not improve risk classification to a statistically significant degree in patients already identified as very low risk by the HEAR score [2].
A Japanese validation by Otsuka and Takeda at a university hospital in Tokyo included 220 patients in the HEAR analysis [4]. Here, however, a different threshold for low risk was used (HEAR 0–3), and the c-statistic for the HEAR score as a continuous variable was 0.76 (95% CI 0.69–0.84) for 6-week MACE. The NPV at the threshold of ≤3 was 0.95 (95% CI 0.89–0.99). The study was small and retrospective, and the results are not directly comparable with the American validations, which use the threshold of ≤1.
Limitations
The HEAR score applies only to patients without known coronary artery disease and without ischaemic ECG changes. In the Ashburn cohort, patients with known coronary artery disease (previous infarction, revascularisation or known stenosis ≥70%) and patients with ischaemic ECG findings (≥1 mm ST depression or T-wave inversion in contiguous leads) were excluded [2]. The HEAR score is not validated in these patients, and troponin testing should be performed whatever the score.
The instrument is not designed to identify high-risk patients. A score of 2 or higher means only that the low-risk strategy cannot be applied and that a full work-up with troponin is required. Interpreting a high HEAR score as a measure of high risk is a misuse.
Because the history component rests on clinical judgement of the character of the symptoms, the score is operator-dependent. Inter-observer variability in assessing the history can affect classification, particularly for patients close to the boundary between 1 and 2 points.
All published validations are from American or Japanese cohorts. There is no Swedish or European validation of the HEAR score at the ≤1 threshold specifically. The original HEART score, on which HEAR is based, is however well validated in European populations.
References
- Moumneh T, Sun BC, Baecker A, et al. Identifying Patients with Low Risk of Acute Coronary Syndrome Without Troponin Testing: Validation of the HEAR Score. Am J Med 2021;134(4):499-506. PMID: 33127371
- Ashburn NP, Snavely AC, Villenthi A, et al. Multisite Validation of a Strategy to Identify Very Low Risk Emergency Department Patients Without Troponin. JACC Advances 2025;4(7). PMID: 40554408
- Yore M, Sharp A, Wu YL, et al. Emergency Department Cardiac Risk Stratification With High-Sensitivity vs Conventional Troponin HEART Pathway. JAMA Network Open 2023;6(12):e2348351. PMID: 38113042
- Otsuka Y, Takeda S. Validation study of the modified HEART and HEAR scores in patients with chest pain who visit the emergency department. Acute Med Surg 2020;7(1):e591. PMID: 33204433
