We sought to evaluate the prognosis performance of the Thrombolysis in Myocardial Infarction (TIMI) risk score in a specific population: ST elevation myocardial infarction (STEMI) patients with reduced left ventricular function.

Among patients included in the randomized EPHESUS trial, we analysed 4304 STEMI patients with left ventricular ejection fraction ≤40%. For each patient, the TIMI risk score were calculated at different time points (30 days, 6 months and 1 year) using specific variables collected.

We confirmed the significant association of TIMI score items with mortality except for diabetes SBP < 100mmHg, weight < 67kg/ and the combined item diabetes / hypertension or angina. As concerning the recurrent myocardial infarction, none of these TIMI risk score items except older patients >75 years were discriminant in multivariate analysis. The univariate analysis of TIMI risk score using a Cox model showed a poor discriminatory accuracy to predict either mortality and recurrent myocardial infarction at different time points attested by c statistic value constantly inferior to 0,65 and 0.66 in overall and revascularized population respectively. After multivariate analysis in subgroup of patients who underwent early revascularization, we demonstrated that LEVF impairment (OR: 1.08 (1.04-1.15, p<0.001), estimated glomerular filtration rate (eGFR) alteration (OR 1.07 (1.1-.1.27), p<0.01) and prior myocardial infarction (OR: 1.67 (1.19-2.37), p=0.004), parameters not included in TIMI risk score, were strongly associated with cardiovascular mortality at 1 year.

TIMI risk score demonstrated a poor discriminatory accuracy in these population. LVEF represents one of the major factors which should not be ignore in a predictive tool evaluating risk score in this situation