TIMI Score for STEMI

Clinical background

The TIMI risk score for STEMI was developed to provide a rapid, bedside risk assessment at presentation in patients with ST-elevation myocardial infarction. The clinical decision is triage and the choice of treatment intensity: which patients can be managed in a standard way, and which require rapidly escalated care, monitoring or transfer to a higher level of care? The score combines eight clinical variables that are immediately available at presentation, unlike more complex models that require laboratory values or echocardiography and therefore cannot be used for acute triage.

Calculating the TIMI risk score

The score is a weighted sum of eight clinical variables, each assigned a value based on its relative contribution to mortality risk in the multivariable model:

$$\text{TIMI} = A + R + B + H + K + W + E + T$$

where the variables and their points are:

The maximum score is 14.

The derivation cohort was the InTIME-II trial, a fibrinolytic trial of 14,114 patients with STEMI eligible for thrombolysis [1]. The mean age was high and mean 30-day mortality in the cohort was 6.7%. The score was modelled against 30-day mortality with logistic regression, and the eight variables together accounted for 97% of the predictive capacity of the full model [1].

Interpretation in practice

In the derivation cohort the score showed a graded rise in mortality from under 1% at a score of 0 to more than a 40-fold increase at a score >8 [1]. Conventionally the score is interpreted in three risk bands:

In a Taiwanese cohort of 796 STEMI patients treated with primary PCI, the median score was 4, and patients with a score $<4$ and no leucocytosis had 0% one-year mortality [3]. This supports the use of a score $<4$ as a threshold for identifying a group at very low risk, even in the PCI era.

Validation and performance

In the derivation cohort the c-statistic was 0.779, comparable with the full multivariable model (0.784) [1]. External validation in the TIMI 9 trial, another fibrinolytic trial, gave a c-statistic of 0.746 [1].

In a PCI-era cohort from Taiwan (796 patients, primary PCI) the c-statistic of the TIMI risk score for one-year mortality was 0.774 [3]. The Chinese Henan-STEMI registry (3,939 patients, primary PCI) found that the TIMI risk score performed equivalently to the GRACE score for 30-day mortality at first medical contact, while an in-hospital stage score including creatinine and left ventricular ejection fraction (LVEF) was superior to both [4]. In a smaller Indian cohort (150 patients, reperfusion with PCI or thrombolysis), the TIMI AUC was 0.74 for mortality and 0.69 for 30-day MACE, while GRACE performed significantly better with an AUC of 0.88 [5].

In a specific subgroup of patients with diabetes and STEMI (455 patients, Taiwan), the TIMI score was inferior to the CADILLAC score, which includes LVEF, renal function and anaemia, for predicting 6-, 12- and 24-month mortality [2]. Calibration for TIMI was, however, good in this cohort (Hosmer–Lemeshow $p = 0{.}40$–$0{.}54$) [2].

Limitations

The fibrinolytic era as derivation context. The score was derived in InTIME-II, in which all patients were treated with fibrinolysis [1]. In the primary PCI era, in which rapid mechanical reperfusion is standard, baseline mortality is lower and the score can be expected to overestimate risk. This is indirectly confirmed by the fact that newer PCI-era scores (CADILLAC, in-hospital stage) perform better than TIMI when PCI data are available [2, 4].

Missing variables. The score does not include renal function, LVEF or cardiac arrest at presentation, variables that in several cohorts have been shown to add predictive value beyond TIMI [2, 3]. In the Taiwanese PCI cohort, adding the white cell count increased the c-statistic from 0.774 to 0.805 [3].

The weight threshold <67 kg. This threshold was derived from a predominantly North American population and reflects the association between low body weight and increased bleeding risk with fibrinolysis. In populations with a different body composition this variable may contribute less to prediction.

Killip class and blood pressure are dynamic. The score is calculated at a single point in time, but the status at presentation can change rapidly. A patient who initially presents in Killip I with a normal blood pressure may quickly develop cardiogenic shock, so the score does not replace continuous clinical observation.

GRACE outperforms TIMI in several cohorts. In direct comparisons, the GRACE score, which includes creatinine and cardiac arrest at presentation, often performs better than TIMI for both mortality and MACE [5]. GRACE is, however, more complex and requires laboratory values, which makes it harder to use for acute triage.

References

1. Morrow DA, Antman EM, Charlesworth A, et al. TIMI risk score for ST-elevation myocardial infarction: a convenient, bedside, clinical score for risk assessment at presentation. Circulation. 2000;102(17):2031–7. PMID: 11044416
2. Kao YT, Hsieh YC, Hsu CY, et al. Comparison of the TIMI, GRACE, PAMI and CADILLAC risk scores for prediction of long-term cardiovascular outcomes in Taiwanese diabetic patients with ST-segment elevation myocardial infarction. PLoS One. 2020;15(2):e0229186. PMID: 32053694
3. Yeh YT, Liu CW, Li AH, et al. Rapid early triage by leukocytosis and the Thrombolysis in Myocardial Infarction (TIMI) risk score for ST-elevation myocardial infarction undergoing primary percutaneous coronary intervention. Medicine (Baltimore). 2016;95(27):e2857. PMID: 26886652
4. Wang S, Zhang Y, Qi D, et al. Development and validation of a risk score for predicting 30-day mortality in patients with ST elevation myocardial infarction. Sci Rep. 2025;15:478255. PMID: 40087520
5. Sr A, Lakshmaiah Y, DGSR KM, et al. A comparison of TIMI and GRACE scores, myocardial blush grade, and shock index-based indices for predicting outcomes in STEMI patients undergoing reperfusion. Cureus. 2025;17(5):e84938. PMID: 40585749