GRACE 2.0 — ACS In-Hospital Mortality Risk
Validated for ACS (STEMI + NSTEMI). Predicts in-hospital mortality risk.
Age (years)
Heart Rate (bpm)
Systolic BP (mmHg)
Creatinine (µmol/L)
Killip Class
Guide détaillé à venir
Nous préparons un guide éducatif complet pour le GRACE ACS Risk Score. Revenez bientôt pour des explications étape par étape, des formules, des exemples concrets et des conseils d'experts.
The GRACE 2.0 ACS Risk Score (Global Registry of Acute Coronary Events) is a validated, multivariate clinical prediction model that quantifies an individual patient's risk of in-hospital and 6-month all-cause mortality following an acute coronary syndrome (ACS). It was derived from the GRACE registry — a large multinational observational study launched in 1999 that enrolled over 102,000 ACS patients across 94 hospitals in 14 countries on 5 continents, making it one of the most geographically diverse and methodologically robust ACS datasets ever assembled. The original GRACE score (2003) was developed by Granger and colleagues using multivariable logistic regression to identify the strongest independent predictors of in-hospital death. The updated GRACE 2.0 model (Eagle et al., 2014) extended predictions beyond the hospital stay to 6-month and 1-year mortality endpoints, and introduced an online nomogram-based approach that does not require complex regression coefficients to be recalled at the bedside. GRACE 2.0 incorporates eight variables: age, heart rate, systolic blood pressure, serum creatinine, Killip class at presentation, whether the patient had a cardiac arrest at admission, ST-segment deviation on ECG, and elevation of cardiac biomarkers (troponin or CK-MB). The score ranges from 0 to 372 and stratifies patients into three mortality risk tiers: low (<108 — predicted 6-month mortality <3%), intermediate (108–140 — predicted 6-month mortality 3–8%), and high (>140 — predicted 6-month mortality >8%). Importantly, the GRACE score was designed specifically for patients with confirmed ACS — both NSTEMI and unstable angina (UA) — unlike the HEART Score, which is an ED triage tool for undifferentiated chest pain. In clinical practice, GRACE is most influential in guiding the timing and urgency of invasive coronary angiography in NSTEMI/UA. The 2020 ESC Guidelines for NSTEMI/UA use GRACE score as a criterion for very high-risk, high-risk, and intermediate-risk stratification, directly determining whether patients should undergo immediate (<2h), early (<24h), or selective invasive strategy. This makes GRACE 2.0 one of the few clinical risk scores that directly maps to a guideline-specified treatment decision, rather than merely serving an advisory function.
GRACE 2.0 Score is derived from logistic regression coefficients applied to 8 clinical variables. The total score ranges 0–372 and is best calculated using the validated online nomogram (gracescore.org) or institutional EHR integration. The 8 variables and their approximate score contributions are: • Age (years): < 30 → 0 pts 30–39 → 8 pts 40–49 → 25 pts 50–59 → 41 pts 60–69 → 58 pts 70–79 → 75 pts 80–89 → 91 pts ≥ 90 → 100 pts • Heart Rate (bpm): < 50 → 0 pts 50–69 → 3 pts 70–89 → 9 pts 90–109 → 15 pts 110–149 → 24 pts 150–199 → 38 pts ≥ 200 → 46 pts • Systolic Blood Pressure (mmHg): < 80 → 58 pts 80–99 → 53 pts 100–119 → 43 pts 120–139 → 34 pts 140–159 → 24 pts 160–199 → 10 pts ≥ 200 → 0 pts (Lower SBP = higher points — inverse relationship) • Creatinine (µmol/L or mg/dL): 0–35 µmol/L (< 0.40 mg/dL) → 1 pt 36–70 µmol/L (0.40–0.79 mg/dL) → 4 pts 71–105 µmol/L (0.80–1.19 mg/dL) → 7 pts 106–140 µmol/L (1.20–1.59 mg/dL) → 10 pts 141–176 µmol/L (1.60–1.99 mg/dL) → 13 pts 177–353 µmol/L (2.00–3.99 mg/dL) → 21 pts > 353 µmol/L (≥ 4.00 mg/dL) → 28 pts • Killip Class: I (no signs of heart failure) → 0 pts II (rales/crackles, S3 gallop, elevated JVP) → 20 pts III (frank pulmonary oedema) → 39 pts IV (cardiogenic shock) → 59 pts • Cardiac Arrest at Admission: No → 0 pts Yes → 39 pts • ST-Segment Deviation: No → 0 pts Yes (new depression or transient elevation in NSTEMI/UA) → 28 pts • Elevated Cardiac Enzymes / Biomarkers (troponin or CK-MB above ULN): No → 0 pts Yes → 14 pts
- 1Confirm the diagnosis of ACS (NSTEMI or unstable angina) based on symptoms, ECG, and initial biomarkers. GRACE 2.0 is validated for use in confirmed ACS patients — it is not an ED triage tool for undifferentiated chest pain. Record the time of symptom onset and first medical contact, as these influence the urgency of scoring.
- 2Record the patient's age in years. Age is the single largest contributor to the GRACE score, reflecting the exponential increase in ACS mortality with advancing age. Each decade after 40 adds approximately 15–17 points. Patients aged ≥ 80 have a baseline GRACE contribution of ~91 points before any other variables are considered.
- 3Measure and document the heart rate on initial presentation. Tachycardia (HR ≥ 110 bpm) reflects sympathetic activation, haemodynamic compromise, or significant left ventricular dysfunction and adds 24–46 points. Bradycardia (HR < 50 bpm) may indicate inferior MI with vagal tone or conduction disease and contributes 0 points but warrants separate clinical attention.
- 4Record the systolic blood pressure (SBP) at first medical contact or ED arrival. Note the inverse scoring relationship: lower SBP contributes MORE points (hypotension = worse prognosis). SBP < 80 mmHg adds 58 points. This captures cardiogenic shock physiology. SBP ≥ 200 mmHg adds 0 points to the score (hypertension alone does not increase GRACE mortality in the acute setting).
- 5Obtain a serum creatinine on the admission blood panel. Renal impairment is an independent predictor of ACS mortality, reflecting both reduced cardiac reserve and limited ability to tolerate interventional procedures and medications. Use local lab units (µmol/L or mg/dL) — convert if needed. Patients with severe CKD (creatinine > 353 µmol/L or 4.0 mg/dL) add 28 points, nearly as much as Killip class II heart failure.
- 6Assess the Killip class at presentation: Class I = no heart failure signs (clear lungs, no S3, no elevated JVP); Class II = mild-to-moderate heart failure (bibasal crackles, S3 gallop, elevated JVP); Class III = acute pulmonary oedema (diffuse rales, severe dyspnoea); Class IV = cardiogenic shock (hypotension SBP < 90, cold extremities, oliguria). The Killip class directly captures acute haemodynamic status and adds 0, 20, 39, or 59 points respectively.
- 7Determine the remaining three binary variables from the initial clinical assessment and investigations: (1) Was there cardiac arrest at the time of admission (VF/pulseless VT requiring resuscitation)? — adds 39 points if yes. (2) Is there ST-segment deviation on the 12-lead ECG (new ST depression ≥ 0.5 mm or transient ST elevation in NSTEMI/UA)? — adds 28 points if yes. (3) Are cardiac biomarkers elevated above the assay upper limit of normal (troponin I/T or CK-MB)? — adds 14 points if yes. Sum all eight components to derive the GRACE 2.0 score and interpret using the risk stratification table.
Low risk — selective invasive strategy acceptable; timing within 72h
This 58-year-old woman with NSTEMI has a GRACE score of approximately 82, driven primarily by age (41 pts), mildly elevated creatinine (7 pts), normal heart rate (9 pts), and biomarker elevation (14 pts). Her preserved blood pressure (SBP 148 adds only ~10 pts), absence of heart failure (Killip I = 0 pts), no cardiac arrest, and no ST deviation keep the score low. Per ESC 2020 NSTEMI guidelines, a GRACE score <109 supports a selective invasive strategy (angiography only if ischaemia confirmed on non-invasive testing or symptoms recur), with outpatient management as an option in selected centres.
Intermediate risk — early invasive strategy recommended within 24h
A 67-year-old male with HR 95, mild renal impairment (creatinine 115 µmol/L), ST depression, and positive troponin accumulates an intermediate GRACE score. Age contributes ~58 pts, HR 90–109 adds 15 pts, SBP 120–139 adds 34 pts, creatinine adds 10 pts, ST deviation adds 28 pts, and biomarkers add 14 pts — totalling approximately 122. This places the patient in the intermediate-risk category. ESC 2020 guidelines recommend early invasive angiography within 24 hours, as this timing has been shown to reduce ischaemic complications compared to delayed or selective approaches in this risk tier.
High risk — early invasive strategy within 24h; ICU-level monitoring required
This 74-year-old with acute pulmonary oedema (Killip III), tachycardia, hypotension, significant renal impairment, anterolateral ST depression, and troponin elevation accumulates a very high GRACE score. Age (~75 pts) + HR 110–149 (24 pts) + SBP 80–99 (53 pts) + creatinine 141–176 µmol/L (13 pts) + Killip III (39 pts) + ST deviation (28 pts) + biomarkers (14 pts) ≈ 246 total. This score well exceeds the high-risk threshold of 140, predicting >8% 6-month mortality. Per ESC 2020, patients with GRACE >140 should receive an early invasive strategy within 24 hours, with ICU or high-dependency unit monitoring.
Very high risk — immediate invasive strategy (<2h); activate cath lab, haemodynamic support
This is the highest-risk clinical phenotype: an 81-year-old with resuscitated cardiac arrest, cardiogenic shock (Killip IV), severe hypotension, significant renal impairment, tachycardia, ST deviation, and markedly elevated biomarkers. Each binary high-risk feature (cardiac arrest +39, Killip IV +59, ST deviation +28, biomarkers +14) alone accounts for 140 additional points. Combined with age ~91 pts, HR +24 pts, SBP <80 = 58 pts, and creatinine = ~21 pts, the total far exceeds 300. Per ESC 2020 NSTEMI guidelines, haemodynamic instability, cardiac arrest, and cardiogenic shock are 'very high risk' criteria mandating immediate invasive strategy within 2 hours, analogous to STEMI treatment.
Guiding the timing of invasive coronary angiography in NSTEMI — ESC 2020 guidelines map GRACE score directly to immediate (<2h), early (<24h), or selective invasive strategy, making it one of the few risk scores with a direct one-to-one mapping to a guideline treatment decision
Prognostic communication with patients and families — a GRACE score >140 corresponds to >8% 6-month mortality, enabling realistic prognostic conversations and advance care planning discussions in high-risk or elderly ACS patients
Guiding intensity of post-discharge medical therapy — high GRACE scores identify patients who benefit most from intensified antithrombotic therapy, high-intensity statin therapy, RAAS inhibition, and close outpatient cardiology follow-up within 2 weeks of discharge
Hospital-level quality benchmarking — GRACE registry data are used by health systems and national cardiac audit programmes (e.g., MINAP in the UK) to compare risk-adjusted ACS outcomes across hospitals and identify centres with above- or below-expected mortality, driving quality improvement initiatives
Clinical trial stratification and outcome adjudication — the GRACE score is used in ACS trial enrolment criteria and subgroup analyses to ensure comparability of risk across randomised arms and to assess treatment effect heterogeneity by baseline risk
GRACE in STEMI — not the primary tool
GRACE 2.0 was derived primarily from NSTEMI and UA patients. In STEMI, immediate primary PCI (within 120 minutes of first medical contact) is the mandated reperfusion strategy regardless of GRACE score — the score does not modify or delay this decision. However, GRACE can be calculated post-procedure in STEMI patients to guide post-discharge risk stratification, intensity of medical therapy, and length of follow-up. The 2020 ESC STEMI guidelines reference GRACE in the context of post-STEMI risk assessment, not acute triage.
Type 2 MI and elevated troponin without obstructive CAD
Type 2 MI (myocardial infarction due to supply-demand mismatch, not plaque rupture) and MINOCA (myocardial infarction with non-obstructive coronary arteries) elevate troponin and may satisfy biomarker criteria for ACS. Applying GRACE to these patients will overestimate obstructive CAD-related mortality risk. Conditions such as myocarditis, Takotsubo syndrome, pulmonary embolism, and severe sepsis cause troponin elevation via non-ACS mechanisms. GRACE should be applied only when type 1 MI (plaque rupture/erosion) is the working diagnosis.
Chronic kidney disease and creatinine interpretation
In patients with pre-existing CKD, the admission creatinine may be chronically elevated and not represent acute deterioration. A creatinine of 250 µmol/L in a patient with known stage 4 CKD adds 21 points to GRACE, reflecting their genuinely higher ACS mortality risk — renal impairment is a true, independent mortality modifier regardless of its acuity. However, clinicians should also document whether the creatinine has changed from baseline (acute-on-chronic kidney injury) and assess contrast risk before angiography. Creatinine clearance (Cockcroft-Gault) rather than absolute creatinine should guide anticoagulant dosing.
Elderly patients and the very high-risk zone
Patients aged ≥ 80 contribute ~91 points from age alone, often placing them in the intermediate-to-high risk range even before other variables are considered. However, very elderly patients also have the highest absolute benefit from invasive revascularisation in ACS — the AFTER-80 trial (Tegn et al., NEJM 2016) demonstrated that patients aged ≥ 80 with NSTEMI had significantly lower rates of MI and need for urgent revascularisation with an invasive strategy. Age alone is not a contraindication to invasive assessment; frailty, comorbidity burden, and patient preference must be incorporated into the shared decision-making process alongside GRACE score.
GRACE in the context of 0h/1h high-sensitivity troponin protocols
Modern high-sensitivity troponin (hs-Tn) assays allow rapid rule-in and rule-out of NSTEMI within 1–2 hours. In centres using ESC 0h/1h or 0h/2h hs-Tn protocols, the biomarker component of GRACE can be determined earlier and with greater precision than with conventional troponin assays. An hs-Tn above the 99th percentile ULN scores the GRACE biomarker criterion as positive (14 pts). Sex-specific hs-Tn thresholds (now recommended by ESC) should be used — the female ULN is typically lower (e.g., hs-TnI < 16 ng/L for men vs < 9 ng/L for women on Abbott ARCHITECT), meaning women may be scored positive at lower absolute troponin levels than men.
| Score Range | Risk Category | 6-Month Mortality | Recommended Action |
|---|---|---|---|
| < 108 | Low | < 3% | Selective invasive strategy; consider non-invasive testing first; outpatient management in stable patients |
| 108–140 | Intermediate | 3–8% | Early invasive strategy within 24–72h; admission for monitoring and serial troponin |
| > 140 | High | > 8% | Early invasive strategy within 24h; ICU or high-dependency monitoring; cardiology input |
| Very high risk features* | Very High | > 15% | Immediate invasive strategy < 2h; equivalent to STEMI protocol; haemodynamic support as needed |
What does the GRACE score predict, and over what time frame?
The GRACE 2.0 score predicts all-cause mortality at two time points: in-hospital mortality (from admission to discharge) and 6-month post-discharge mortality. The score can also estimate 1-year mortality using the GRACE 2.0 nomogram. In clinical use, the 6-month mortality prediction is most commonly used for risk stratification and treatment planning. The score does not directly predict non-fatal MI, revascularisation, or rehospitalisation, although higher-risk patients are at greater risk of all adverse cardiovascular events.
Who should the GRACE score be applied to?
GRACE 2.0 is validated for patients with confirmed ACS — specifically NSTEMI (non-ST-elevation myocardial infarction) and unstable angina (UA). It was derived from a registry of hospitalised ACS patients and is NOT intended for: (1) ED triage of undifferentiated chest pain (use HEART Score instead), (2) STEMI patients (who require immediate reperfusion regardless of score), or (3) patients with non-ACS causes of troponin elevation such as myocarditis, pulmonary embolism, or type 2 MI. Applying GRACE to unconfirmed or non-ACS populations will give misleading results.
How does GRACE 2.0 differ from the original GRACE score?
The original GRACE score (published 2003) used the same eight variables to predict in-hospital mortality in ACS. GRACE 2.0 (Eagle et al., 2014) extended the model to predict post-discharge outcomes at 6 months and 1 year, which is clinically more important for guiding discharge planning and long-term medical therapy. GRACE 2.0 also introduced a simplified online calculator and nomogram approach (gracescore.org) that does not require manual coefficient lookup tables, improving bedside usability. The risk thresholds (low <108, intermediate 108–140, high >140) refer specifically to the GRACE 2.0 6-month mortality model.
How does GRACE guide the timing of coronary angiography in NSTEMI?
The 2020 ESC NSTEMI/UA Guidelines directly incorporate GRACE score into the invasive strategy timing decision: (1) Very high risk features (haemodynamic instability, cardiac arrest, Killip IV, mechanical complications) → immediate invasive (<2h) regardless of GRACE score; (2) High risk (GRACE >140, or dynamic troponin, or ST changes) → early invasive within 24h; (3) Intermediate risk (GRACE 109–140, diabetes, eGFR <60, LVEF <40%, post-MI angina) → invasive within 72h; (4) Low risk (GRACE <109, no high-risk features, stable with negative biomarkers) → selective invasive (non-invasive testing first). This clear mapping of GRACE to treatment decision timing makes it uniquely actionable among ACS risk scores.
What is Killip class and why does it contribute so heavily to the GRACE score?
Killip classification describes the degree of haemodynamic compromise from left ventricular failure in ACS: Class I (no heart failure signs, estimated in-hospital mortality 5%), Class II (mild-moderate HF with rales, S3 gallop, elevated JVP; mortality ~12%), Class III (acute pulmonary oedema; mortality ~40%), Class IV (cardiogenic shock; mortality ~55–80%). Cardiogenic shock complicates 5–8% of NSTEMI cases but accounts for >60% of in-hospital deaths. The Killip classification contributes 0, 20, 39, or 59 points respectively in GRACE, making it the largest single modifiable contributor to the score — reflecting that acute haemodynamic status is the primary determinant of short-term mortality after the infarct has occurred.
How does renal function affect the GRACE score and clinical management?
Renal impairment is an independent predictor of ACS mortality through multiple mechanisms: reduced clearance of cardiotoxic metabolites, renin-angiotensin system activation, accelerated atherosclerosis, and impaired platelet function. Patients with severe CKD (creatinine >353 µmol/L or 4.0 mg/dL) add 28 points to GRACE — nearly as much as acute pulmonary oedema (Killip III adds 39 points). Beyond the score, renal function critically influences medication dosing: anticoagulants such as enoxaparin, bivalirudin, and fondaparinux all require dose adjustment or avoidance in severe CKD; GPI agents (eptifibatide, tirofiban) accumulate in renal failure; and contrast-induced nephropathy risk must be minimised during angiography with adequate pre-hydration.
What is the significance of cardiac arrest at admission in the GRACE score?
Cardiac arrest at admission (defined as resuscitated VF or pulseless VT requiring CPR or defibrillation before or at the time of hospitalisation) adds 39 points to GRACE, equivalent to Killip class II-to-III transition, reflecting the profound prognostic weight of pre-hospital cardiac arrest. Importantly, this variable specifically captures PRIMARY ventricular arrhythmia as an ACS complication, not respiratory arrest or shock from other causes. In the ESC 2020 guidelines, out-of-hospital cardiac arrest is itself a very high-risk criterion triggering immediate invasive strategy within 2 hours, effectively superseding the calculated GRACE score as the primary triage criterion in that clinical scenario.
Can the GRACE score be used in patients who cannot provide a full history (e.g., post-arrest, intubated)?
Yes, with important nuances. The GRACE score relies on objective clinical and laboratory data that can be obtained even from obtunded or intubated patients: age (from records), heart rate and SBP (from monitoring), creatinine (from lab), Killip class (from clinical examination and CXR), cardiac arrest status (from EMS/witness report), ECG findings, and biomarkers. However, in post-cardiac arrest patients or those with haemodynamic instability, the clinical situation itself (Killip IV, cardiac arrest) already mandates immediate invasive strategy per guideline very-high-risk criteria — meaning the GRACE score calculation, while still informative, is not the primary treatment trigger in these extreme cases.
Conseil Pro
Always cross-reference the calculated GRACE score with the ESC 2020 'very high risk' criteria checklist — haemodynamic instability, cardiac arrest, acute heart failure, mechanical complications, and life-threatening arrhythmias mandate immediate invasive strategy (<2h) regardless of score. In practice, use the validated online GRACE 2.0 calculator at gracescore.org or your institutional EHR integration rather than manually summing point tables, as regression-based predictions are more accurate than simplified lookup tables. Document the score in the medical record at admission and reassess if haemodynamic status changes significantly during the first 12–24 hours.
Le saviez-vous?
The GRACE registry was one of the first truly global cardiovascular registries, enrolling patients across 14 countries — from Argentina to Australia — when it launched in 1999. It captured ACS management during the era before widespread high-sensitivity troponin, radial-access PCI, and potent P2Y12 inhibitors. The finding that in-hospital outcomes varied dramatically by country (despite similar patient demographics) drove major international efforts to standardise ACS care pathways. Today, GRACE-derived risk data underpin multiple international ACS guideline recommendations, and the registry continues to enrol patients, providing longitudinal data spanning more than two decades of cardiology practice evolution.
Références
- ›Granger CB et al. Predictors of hospital mortality in the Global Registry of Acute Coronary Events — Arch Intern Med 2003
- ›Eagle KA et al. A validated prediction model for all forms of acute coronary syndrome: estimating the risk of 6-month postdischarge death in an international registry (GRACE 2.0) — Eur Heart J 2004
- ›Collet JP et al. 2020 ESC Guidelines for the management of acute coronary syndromes in patients presenting without persistent ST-segment elevation — Eur Heart J 2021
- ›Fox KA et al. Should patients with acute coronary disease be stratified for management according to their risk? Derivation, external validation and outcomes using the updated GRACE risk score — BMJ Open 2014
- ›Tegn N et al. Invasive versus conservative strategy in patients aged 80 years or older with non-ST-elevation myocardial infarction or unstable angina pectoris (After Eighty study) — Lancet 2016