Prognostic performance of GRACE and TIMI risk scores in critically ill patients with sepsis and a concomitant myocardial infarction

Abstract

Identifying which patients with acute myocardial infarction (AMI) during sepsis are at risk of poor outcome is a clinical challenge. To evaluate Global Registry of Acute Coronary Events (GRACE) and Thrombolysis In Myocardial Infarction (TIMI) risk scores to predict in-hospital mortality and severe ischaemic events in this setting. In this single-centre retrospective study conducted from 2012 to 2016, all consecutive adults hospitalized in the intensive care unit for sepsis who had a concomitant AMI (within 72hours of admission) were enrolled. AMI was defined by an elevated cardiac troponin I value associated with at least one sign (clinical, electrocardiographic or echocardiographic) suggestive of myocardial ischaemia. The primary outcome was in-hospital mortality from any cause. Secondary outcomes were in-hospital occurrence of severe ischaemic events (cardiac arrest with resuscitation, ischaemic stroke and myocardial reinfarction) and major bleeding events. Among 856 patients hospitalized for sepsis, 120 (14.5%) had a concomitant AMI (37.5% women; median age 65 years; median Sequential Organ Failure Assessment [SOFA] score 8). Severe ischaemic events occurred in 15 patients (12.5%), and 39 (33%) died in hospital. Neither the GRACE score (median 192, interquartile range 154-223) nor the TIMI score (median 3, interquartile range 2-4) was associated with occurrence of severe ischaemic events. Only the GRACE score was associated with in-hospital mortality (odds ratio 1.01, 95% confidence interval 1.00-1.02 per 1 point increase). Multivariable analysis identified previous aspirin use and SOFA score as independent factors associated with in-hospital mortality. GRACE and TIMI scores did not predict in-hospital severe ischaemic events and mortality in patients with AMI during sepsis. Among individual components of both scores, previous aspirin use was associated with poor prognosis. However, because of lack of statistical power, we cannot formally rule out the usefulness of these scores in this setting.