Chest pain risk scores

HEART was built for undifferentiated ED chest pain; TIMI and GRACE were derived in confirmed ACS.

Side-by-side comparison of model characteristics
Characteristic	HEART Score	TIMI (UA/NSTEMI)	GRACE (in-hospital)
Clinical question	What is the short-term risk of a major adverse cardiac event in an adult with chest pain in the ED?	What is the 14-day risk of death or ischaemic events in confirmed UA/NSTEMI?	What is the risk of death during hospitalisation for acute coronary syndrome?
Outcome predicted	Major adverse cardiac events (death, MI, coronary revascularisation)	All-cause mortality, new or recurrent MI, or severe recurrent ischaemia requiring urgent revascularisation	In-hospital death
Time horizon	6 weeks	14 days	Index hospitalisation
Intended population	Adult ED patients with chest pain in Dutch hospitals (derivation 2008; multicentre validation 2013).	Patients with UA/NSTEMI enrolled in the TIMI 11B and ESSENCE trials.	Patients with ACS in the multinational Global Registry of Acute Coronary Events.
Evidence level	Strong evidence	Strong evidence	Strong evidence
Validation	Prospective multicentre validation (Backus 2013): 6-week MACE: low 1.7%, moderate 16.6%, high 50.1%; C-statistic 0.83.	Derivation (TIMI 11B): 14-day event rates by score: 0–1, 4.7%; 2, 8.3%; 3, 13.2%; 4, 19.9%; 5, 26.2%; 6–7, 40.9%.	—
Discrimination	Reported C-statistic 0.83 in the 2013 multicentre validation (vs 0.75 TIMI and 0.70 GRACE in the same cohort).	—	C-statistic 0.83 (derivation), 0.84 (GRACE confirmation cohort) and 0.79 (GUSTO-IIb external cohort).
Calibration	—	—	—
Geographic applicability	Validated in Europe, North America and Asia.	—	Multinational registry (14 countries).
Guideline status	Listed among clinical decision pathways in the 2021 AHA/ACC multi-society chest pain guideline.	Historically referenced in ACC/AHA NSTE-ACS guidance for risk stratification.	Recommended for prognostic risk stratification in ESC acute coronary syndrome guidelines.
Important limitations	History and ECG items are subjective, with moderate inter-rater agreement. Developed with conventional troponin; high-sensitivity assays require validated accelerated pathways. Not a substitute for serial troponin testing.	Derived in trial populations with confirmed ACS — less discriminating in undifferentiated ED chest pain. Aspirin use as a risk factor reflects the derivation era.	Derived in an earlier registry era; contemporary treatment may lower absolute risks. This is GRACE v1.0; GRACE 2.0 (non-linear, 1- and 3-year outcomes) is not implemented. Killip class and creatinine are required for this version.

Inputs required

Which variables each model uses. Models with more inputs are not automatically better — validation in your population matters most.

Input variables used by each model
Input	HEART Score	TIMI (UA/NSTEMI)	GRACE (in-hospital)
History	●used	–not used	–not used
ECG	●used	–not used	–not used
Age	●used	–not used	●used
Risk factors	●used	–not used	–not used
Initial troponin	●used	–not used	–not used
Age ≥ 65 years	–not used	●used	–not used
≥ 3 CAD risk factors	–not used	●used	–not used
Known CAD (stenosis ≥ 50%)	–not used	●used	–not used
Aspirin use in the past 7 days	–not used	●used	–not used
Severe angina (≥ 2 episodes in 24 h)	–not used	●used	–not used
ST deviation ≥ 0.5 mm on presenting ECG	–not used	●used	–not used
Positive cardiac marker	–not used	●used	–not used
Heart rate	–not used	–not used	●used
Systolic BP	–not used	–not used	●used
Creatinine	–not used	–not used	●used
Killip class	–not used	–not used	●used
Cardiac arrest	–not used	–not used	●used
ST-segment deviation	–not used	–not used	●used
Elevated cardiac biomarkers	–not used	–not used	●used

Comparison shows model characteristics only. Different models predict different outcomes in different populations, so their numerical results are not directly comparable.