HEART Score for chest pain

Clinical background

The HEART score is intended to risk-stratify adult patients presenting to the emergency department with chest pain in whom acute coronary disease is suspected but the picture is not obvious. The decision the tool serves is one of triage: which patients can be discharged with reasonable safety, which need admission for observation and further investigation, and which should proceed to early invasive investigation? Without a structured tool, clinicians tend either to over-read non-specific chest pain and admit patients unnecessarily, or to minimise symptoms and miss patients with early acute coronary syndrome. The HEART score is designed specifically for the large intermediate group of patients with undifferentiated chest pain, in whom the ECG does not show STEMI and the initial troponin is not unequivocally raised.

Calculating the HEART score

$$\text{HEART} = H_{\text{history}} + E_{\text{ECG}} + A_{\text{age}} + R_{\text{risk factors}} + T_{\text{troponin}}$$

Each variable is assigned 0, 1 or 2 points as follows:

The risk factors counted are hypertension, hypercholesterolaemia, diabetes, obesity, smoking and family history. The maximum score is 10.

The derivation cohort consisted of 122 patients (120 of whom were analysed) presenting to a Dutch emergency department with chest pain [1]. The outcome measure was a composite of acute myocardial infarction, revascularisation or death. Of these 120 patients, 29 reached an outcome. The score differed significantly between patients with and without an outcome (6.51 and 3.71 respectively, p<0.0001) [1]. The cohort was small and single-centre, which makes external validation essential in judging the reliability of the tool.

Interpretation in practice

The score is divided into three risk bands with clear consequences for management:

In the derivation cohort the risk of an outcome was 2.5% at a score of 0–3, 20.3% at 4–6 and 72.7% at ≥7 [1]. In the large external validation cohort (see below) the corresponding figures were 1.7%, 16.6% and 50.1% [2]. The low band has thus been confirmed as safe for discharge in several independent cohorts, while the high band warrants aggressive management. The intermediate band is the most difficult: here practice varies between admission and rapid outpatient follow-up, and a retrospective study from Michigan showed that selected patients with a HEART score of 4–6 could be discharged with cardiology follow-up within approximately 3 working days, with a MACE rate below 1% at 30 days [3]. This does, however, require structured follow-up pathways and shared decision-making with the patient.

Validation and performance

The most important external validation was carried out by Backus et al. in a prospective multicentre study at ten Dutch hospitals, comprising 2,440 unselected chest pain patients [2]. The primary outcome was MACE within 6 weeks. The c-statistic for HEART was 0.83, significantly higher than for TIMI (0.75) and GRACE (0.70) (p<0.0001) [2]. At the threshold of ≤3 points, which applied to 36.4% of patients, the MACE rate was 1.7%, meaning that short-term MACE could be excluded with >98% certainty [2].

A systematic review and meta-analysis of 33 prospective cohort studies with a total of 40,262 patients found a pooled sensitivity of 0.96 and a negative likelihood ratio of 0.08 for HEART at the threshold of ≤3 [4]. The AUC was 0.80 (95% CI 0.77–0.84) [4]. HEART performed comparably to TIMI and better than GRACE in both discrimination and diagnostic odds ratio [4]. Heterogeneity was, however, substantial (I² up to 99%), reflecting variation in populations, troponin assays and MACE definitions between studies.

In an Iranian external validation of 274 patients at two cardiology emergency departments, HEART achieved an AUC of 0.925 and a sensitivity of 97.1% at the threshold of ≤3, with a negative predictive value of 98.2% [5]. This is higher than in most Western cohorts and may reflect a population with a higher pre-test probability at a specialist cardiology emergency department.

Limitations

The HEART score does not apply to patients with STEMI on the ECG or with haemodynamic instability. These patients should be managed according to ACS protocols whatever the score. Nor is the score validated for patients with non-cardiac chest pain as the primary working diagnosis, such as pulmonary embolism or aortic dissection.

The most obvious weakness is the troponin variable. The score rests on a threshold that multiplies the laboratory's upper limit of normal by 1 and by 3, but different troponin assays have different limits of normal and different dynamics at early presentation. With the move to high-sensitivity troponin assays the pattern has changed: a patient may have a rise that scores 2 points on HEART but would have been undetectable with older assays. This means that the calibration of the score is assay-dependent and may differ between laboratories.

The history variable is subjective and has proved to be the component with the lowest inter-observer reliability. In the Iranian validation the history was classified by a junior doctor, which may have affected the results [5]. In clinical practice the judgement of "moderately suspicious" versus "highly suspicious" varies between assessors, and this contributes to the spread in scoring.

The derivation cohort was small (120 patients) and single-centre [1]. That the tool nonetheless performs well in external validations reflects the fact that the variables are clinically sensible and the scoring simple, but it also means that the exact risk percentages in each band should be read as indicative rather than as precise estimates.

References

1. Six AJ, Backus BE, Kelder JC. Chest pain in the emergency room: value of the HEART score. Neth Heart J. 2008;16(6):191–6. PMID: 18665203
2. Backus BE, Six AJ, Kelder JC, et al. A prospective validation of the HEART score for chest pain patients at the emergency department. Int J Cardiol. 2013;168(3):2153–8. PMID: 23465250
3. Pawlikowski A, Hubbard E, Krauss J, et al. Early emergency department discharge for intermediate heart score patients presenting for chest pain. J Am Coll Emerg Physicians Open. 2023;4(5):e13037. PMID: 37692195
4. Ke J, Chen Y, Wang X, et al. Indirect comparison of TIMI, HEART and GRACE for predicting major cardiovascular events in patients admitted to the emergency department with acute chest pain: a systematic review and meta-analysis. BMJ Open. 2021;11(8):e048356. PMID: 34408048
5. Nasr Isfahani M, Mohseni H, Nasri Nasrabadi E, et al. Improving chest pain risk assessment: validation of HEART, TIMI, GRACE, EDACS-ADP, and HET for MACE prediction in the emergency department. BMC Emerg Med. 2025;25:165. PMID: 40847270