INTRODUCTION — In patients with ST-elevation myocardial infarction (STEMI), the likelihood of adverse events can be estimated from clinical assessment, risk factors, and risk models.
The general approach to risk stratification for patients with STEMI will be reviewed here.
Risk stratification for patients with acute non-ST elevation acute coronary syndromes and for those at risk for life-threatening arrhythmias is discussed separately. (See "Risk stratification after non-ST elevation acute coronary syndrome" and "Incidence of and risk stratification for sudden cardiac death after myocardial infarction".)
EARLY RISK STRATIFICATION — In general, risk stratification does not influence the acute management of patients with STEMI. Patients with STEMI can undergo risk assessment with either the TIMI risk score (calculator 1) or the GRACE risk model (table 1). These tools include predictors of poor outcomes identified in large databases of patients with STEMI. (See "Risk factors for adverse outcomes after ST-elevation myocardial infarction".)
A report published in 1998 from the National Registry of Myocardial Infarction (NRMI) evaluated data on 170,143 patients admitted with an acute MI (with or without ST-segment elevation) in an attempt to identify patients at high risk [1]. Significant risk factors included age over 70 years, prior MI, Killip class at admission (table 2), anterior MI, and the combination of hypotension and tachycardia.
Similar data were reported by the GUSTO trial of 41,021 patients, which found that predictors of one-year mortality among those who survived to 30 days after their MI included [2]:
●Demographics, including older age (>55), lower weight (≤80 kg), previous MI, and previous bypass surgery. In another population-based study of 2541 patients, severe obesity (BMI >30 kg/m2) also increased the risk of a cardiac event [3]. (See "Overweight and obesity in adults: Health consequences", section on 'Heart disease'.)
●Larger MIs, as determined by higher Killip class, anterior wall MI, lower blood pressure, faster heart rate (>115 bpm), longer QRS duration (>125 ms), lower left ventricular ejection fraction (LVEF), heart failure (HF) and pulmonary edema, and cardiogenic shock.
●Presence of cardiac risk factors, including smoking, hypertension, and prior cerebrovascular disease.
●Other findings such as in-hospital stroke, ventricular or supraventricular arrhythmias, absence of revascularization, and being a Black person.
Based upon these findings and analyses from other large clinical trials and registries, a number of different risk scores have been developed to assess short- and long-term outcomes after STEMI [4-10]. Only two specifically addressed patients treated with primary percutaneous coronary intervention, which is the current modality of choice [5,7].
TIMI risk score — The TIMI risk score, based upon data from 15,000 patients with an STEMI eligible for fibrinolytic therapy, is an arithmetic sum of eight independent predictors of mortality [4]:
●Age ≥75 years – 3 points
●Age 65 to 74 years – 2 points
●History of diabetes, hypertension, or angina – 1 point
●Systolic blood pressure <100 mmHg – 3 points
●Heart rate >100/min – 2 points
●Killip class II to IV (table 2) and – 2 points
●Weight <67 kg – 1 point
●Anterior ST elevation or left bundle branch block – 1 point
●Time to reperfusion therapy >4 hours – 1 point
There is a continuous relationship between mortality and score; a score of 0 to >8 was associated with a 30-day mortality of 0.8 to 36 percent, while the one-year mortality among those surviving the first 30 days ranges from 1 to 17 percent (calculator 1).
The accuracy of the TIMI risk score to predict in-hospital mortality was validated in a community-based population of 84,029 patients; the predictive accuracy of the TIMI risk score was the same in those treated with fibrinolysis or percutaneous coronary intervention, but underestimated mortality in those not undergoing reperfusion therapy (calculator 1) [11].
TIMI risk index — The TIMI risk index (TRI) is a simpler model derived from the InTIME-II trial of fibrinolytic therapy [8] and then validated in other populations to predict in-hospital mortality [12,13]. It can be used simultaneously with the TIMI risk score
The TRI is calculated from the following equation, using data obtained at presentation (table 3) [8]:
TRI = (Heart rate x [age/10] squared) / systolic pressure
The TRI was applied to over 153,000 patients with STEMI in the National Registry of Myocardial Infarction (NRMI) in the United States [12]. There was a graded relationship to in-hospital mortality, ranging from 0.6 to 60 percent from the lowest (0 to <10) to the highest scores (≥80) in patients who received reperfusion therapy and from 1.9 to 52.2 percent in patients who did not receive reperfusion therapy [12]. Patients with a TRI <30 were at low risk. The percent of patients at high risk (TRI >60) was 14.1 percent in those who were not reperfused compared to 2.1 percent in those who were reperfused [14].
GRACE risk model — The TIMI risk score was derived from clinical trial databases, although it has been validated in community-based populations [11,12]. The GRACE registry, a global registry of acute coronary syndrome (ACS) patients from 94 hospitals in 14 countries, developed two models to estimate the risk of both in-hospital and six-month mortality among all patients with an ACS.
The in-hospital model was based upon data from 11,389 patients with either an STEMI or a non-ST elevation ACS [5]. This model was then validated based upon data from an additional 3972 patients from GRACE and 12,142 patients from the GUSTO IIb trial. Eight independent risk factors were found to account for almost 90 percent of the prognostic information:
●Age
●Killip class (table 2)
●Systolic blood pressure
●Presence of ST-segment deviation
●Cardiac arrest during presentation
●Serum creatinine concentration
●Presence of elevated serum cardiac biomarkers
●Heart rate
Point scores were assigned for each predictive factor and are added together to arrive at an estimate of the risk of in-hospital mortality. A nomogram was published with the GRACE risk model to allow calculation of the risk score [5,15].
The six-month model was based upon data from 15,007 patients and validated in a cohort of 7638 patients, all in the GRACE registry [15]. The variables incorporated into this model include age, prior history of HF, prior history of MI, resting heart rate, systolic blood pressure, ST-segment depression, initial serum creatinine concentration, elevated serum cardiac biomarkers, and performance of in-hospital percutaneous coronary intervention (PCI). The six-month mortality risk based upon this model can be calculated using a website.
CHADS2 score — While the GRACE prediction model is well validated and its use is recommended by multiple guideline organizations, its complexity makes it somewhat difficult to use in some clinical settings. The value of the CHADS2-VASc score, which is a well validated tool for predicting the risk of stroke in patients with atrial fibrillation, was evaluated in a study of more than 2300 patients with ACS (37 percent with STEMI; 19 percent with atrial fibrillation [AF]) cared for between 1995 and 2001 [16]. All-cause mortality at 10 years was strongly associated with the CHADS2 score in patients with and without AF. As expected, the more complex GRACE score provided a better prediction for short- and long-term mortality. (See "Atrial fibrillation in adults: Selection of candidates for anticoagulation", section on 'Use'.)
ACTION registry score — The Acute Coronary Treatment and Intervention Outcomes Network (ACTION) Registry was used to develop a risk score to predict in-hospital mortality following STEMI and NSTEMI [17]. Multivariable analyses of data from 243,440 patients showed that greater heart rate, lower systolic blood pressure, life-threatening presentations (cardiac arrest, cardiogenic shock, or HF), STEMI at presentation, lower creatinine clearance, and higher troponin values were associated with death during the hospitalization. The C statistic was very good at 0.88.
The ACTION score is likely to be most helpful for persons with moderate to severe disease and reflects more contemporary experience than TIMI or GRACE.
Two additional multivariable models have been devised and validated for patients exclusively undergoing primary PCI: the Zwolle primary PCI risk index and the CADILLAC risk score.
Zwolle primary PCI index — A risk index based upon a primary PCI population was developed in Zwolle, the Netherlands from data on 1791 patients undergoing primary PCI between 1994 and 2001 [6]. Significant independent risk factors for 30-day mortality were incorporated into the Zwolle index, including Killip class (table 2) and post-PCI TIMI flow grade (table 4), age, number of diseased vessels, location of infarction, and time to reperfusion. The risk index was validated in an additional group of 747 patients with similar characteristics treated with primary PCI between 2001 and 2003.
Based upon the Zwolle risk index, more than two-thirds of patients undergoing primary PCI were classified as low risk (risk score ≤3) [6]. For these patients, the mortality rate was 0.1 percent at two days and 0.2 percent between 2 and 10 days post-MI. It was suggested that such low-risk patients could safely be discharged early (48 hours after PCI).
CADILLAC PCI risk score — A second primary PCI risk model was derived from the 2082 patients in the CADILLAC trial of abciximab or placebo and stenting or angioplasty in primary PCI and then validated using data from the 900 patients in the Stent-PAMI trial [7]. Seven variables, which are readily available at the time of intervention, were weighted according to their odds ratio for one-year mortality (table 5):
●LVEF <40 percent – 4 points
●Killip class 2/3 – 3 points
●Renal insufficiency (estimated creatinine clearance <60 mL/min) – 3 points
●TIMI flow grade after PCI 0 to 2 – 2 points
●Age >65 years – 2 points
●Anemia (hematocrit <39 percent in men and <36 percent in women) – 2 points
●Triple-vessel disease – 2 points
In both the derivation and validation models, patients could be stratified into three risk groups that predicted 30-day and one-year mortality:
●Low risk (score 0 to 2) – 0.1 to 0.2 percent at 30 days and 0.8 to 0.9 percent at one year
●Intermediate risk (score 3 to 5) – 1.3 to 1.9 percent at 30 days and 4.0 to 4.5 percent at one year
●High risk (score ≥6) – 6.6 to 8.1 percent at 30 days and 12.4 to 13.2 percent at one year
The percentage of patients in these three groups was 56, 24, and 20 percent, respectively. The results compared favorably with the TIMI risk score, which was based upon patients undergoing fibrinolysis, and Zwolle primary PCI index.
Comparison of risk scores — The prognostic value of the TIMI, PAMI, CADILLAC PCI, and GRACE risk scores was directly compared in 855 registry patients with STEMI, but without cardiogenic shock, who underwent primary PCI [18]. The TIMI, PAMI, and CADILLAC scores had relatively high and similar predictive accuracies for 30-day and one-year mortality; the GRACE model performed less well. The CADILLAC PCI score was judged to be unhelpful prior to angiography because LVEF is a key component.
Machine learning models — Machine learning models that predict in-hospital mortality after acute MI have been created, but are not more accurate than the models listed above [19].
LATE RISK STRATIFICATION — The major components of late risk stratification are clinical status and LVEF. Methods used to determine the risk of arrhythmic death are discussed separately. These include ventricular arrhythmias and late potentials on signal-averaged electrocardiogram (ECG). (See "Incidence of and risk stratification for sudden cardiac death after myocardial infarction".)
Left ventricular ejection fraction — Assessment of resting LV function is an important part of risk stratification in patients with acute MI and was recommended by the American College of Cardiology Foundation/American Heart Association (ACCF/AHA) and 2012 European Society of Cardiology STEMI guidelines [20-22]. Echocardiography is preferred to magnetic resonance imaging in the European guideline, while the United States guideline makes no preference for the type of imaging modality. Patients with LV systolic dysfunction, defined as LVEF <50 percent, have increased mortality at six months and one year (figure 1 and figure 2) [23-25]. The increase in mortality is most pronounced in the minority of patients with an LVEF ≤30 percent. In addition, patients with LVEF ≤35 percent are at increased risk for sudden cardiac death after MI and should be considered candidates for an implantable cardioverter-defibrillator. Moreover, the degree of LVEF between two and seven days following an acute MI and second determination 2 to 12 weeks following an MI predicted a high risk of sudden death and mortality. Patients with no recovery of LVEF between these two time intervals are at a particularly high risk for these complications [26]. (See "Primary prevention of sudden cardiac death in patients with cardiomyopathy and heart failure with reduced LVEF".)
LVEF is usually measured before discharge in the absence of a specific indication (eg, HF or suspected mechanical complication) for which echocardiography may be indicated early in the hospitalization. However, measurements during hospitalization may be misleading, since improvement in LVEF, beginning within three days and largely complete by 30 days, is common in patients who are reperfused [27]. Two separate studies have shown that approximately 58 percent of patients significantly improve their LVEF after reperfusion in acute STEMI [28,29]. This may reflect, at least in part, recovery from myocardial stunning [28,30] since it is associated with a reduction in the size of the myocardial perfusion defect [29]. Patients with improved LVEF may have significantly lower mortality than those who show no improvement (1.2 versus 5.6 percent at three years in one study) [29]. (See "Clinical syndromes of stunned or hibernating myocardium", section on 'Acute myocardial infarction'.)
Among patients with an STEMI who have an interpretable ECG (ie, no left bundle branch block, paced rhythm, or LV hypertrophy with strain pattern), the absence of an anterior infarction, a previous Q wave MI, or a history of HF predicts an LVEF of at least 40 percent (positive predictive value of 91 to 98 percent) [31,32]. The LVEF is variable among patients who do not fit the prediction rule.
Right ventricular ejection fraction — Long-term impairment of right ventricular systolic function after MI is associated with a worse prognosis (see "Right ventricular myocardial infarction", section on 'Long-term prognosis'). One study noted that patients with STEMI who had right ventricular remodeling (dilation) had a significantly lower survival rate [33].
Mitral regurgitation — In addition to the poor prognosis associated with acute mitral regurgitation (MR) caused by papillary muscle dysfunction or rupture, chronic MR is also associated with a worse prognosis:
●In one study, moderate to severe ischemic MR was present in 9.3 percent of patients, and one-year mortality was higher in those with MR compared with those who did not have MR after MI (15 versus 7.3 percent; adjusted HR 1.4, 95% CI 1.1-2.2) [34].
●In another study of patients who had an MI, moderate to severe MR was present in 25 percent of patients, and survival at one year was lower in patients with moderate to severe MR compared with those without MR (approximate one-year survival (80 versus 95 percent) [35].
Additional information on acute and chronic MR can be found separately. (See "Acute mitral regurgitation in adults" and "Clinical manifestations and diagnosis of chronic mitral regurgitation", section on 'Identifying the cause of MR'.)
Stress testing — The role of noninvasive stress testing after STEMI is discussed elsewhere. (See "Overview of the nonacute management of ST-elevation myocardial infarction", section on 'Stress testing'.)
Other post-myocardial infarction risk factors — A variety of other characteristics have been shown to correlate with late prognosis following acute MI. For example, the extent of angiographic coronary artery disease was a predictor for the development of HF [36]. Environmental factors such as living in an area with a lower mean income or being socially disadvantaged increased the risk for post-MI hospitalization and mortality [37-39]. Two biomarkers, N-terminal pro-B-type natriuretic peptide (NT-proBNP) and growth differentiation factor-15, were strongly associated with all-cause death in patients with an acute coronary syndrome, most of whom had an acute MI [40]. Worsening postdischarge renal function in patients with type 2 diabetes mellitus and a recent ACS were strong predictors of adverse cardiovascular events, including all-cause mortality [41]. Additionally, patients who had influenza and other viral respiratory infections concomitant with acute MI had worse outcomes than patients without viral syndromes [42]. Another research group developed a 30-day and one-year risk model for predicting mortality in ACS. They used a variety of biomarkers including troponin, creatinine clearance, NT-proBNP, C-reactive protein, and age in their model, which had a high predictive value [43].
Young adults with myocardial infarction — Over a median follow-up of more than 11 years, very young patients (<40 years of age) with MI had a similar risk of all-cause and cardiovascular mortality when compared with patients ages 41 to 50 [44].
Sex differences in cardiac death after myocardial infarction — Females with STEMI have been noted to have worse outcomes compared with males when adjusted for the anatomic complexity of CAD. This is particularly true for females who undergo interventional therapy for STEMI [45,46].
SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: ST-elevation myocardial infarction (STEMI)".)
SUMMARY AND RECOMMENDATIONS
●Timing of stratification – In patients with ST-elevation myocardial infarction (STEMI), the likelihood of adverse events can be estimated from clinical assessment, risk factors, and risk models.
•Early risk stratification – In general, risk stratification does not influence the acute management of patients with STEMI. Patients with STEMI can undergo risk assessment with either the TIMI risk score (calculator 1) or the GRACE risk model (table 1). (See 'TIMI risk score' above and 'GRACE risk model' above.)
•Late risk stratification – The major components of late risk stratification are clinical status and left ventricular ejection fraction (LVEF). (See 'Left ventricular ejection fraction' above.)
●Risk of sudden death – Methods used to determine the risk of arrhythmic death are discussed separately. These include ventricular arrhythmias and late potentials on signal-averaged ECG. (See "Incidence of and risk stratification for sudden cardiac death after myocardial infarction".)
