OUP user menu

★ Editor's choice ★

Impact of ESC/ACCF/AHA/WHF universal definition of myocardial infarction on mortality at 10 years

Francisco Moscoso Costa, Jorge Ferreira, Carlos Aguiar, Hélder Dores, João Figueira, Miguel Mendes
DOI: http://dx.doi.org/10.1093/eurheartj/ehs311 2544-2550 First published online: 18 September 2012

Abstract

Aims Redefinition of myocardial infarction (MI) based on specific cardiac troponins (cTn) was universally accepted in 2007. The new definition has been widely discussed for including a large spectrum of quantitative myocardial necrosis and their clinical implications remain under debate. Our aim was to assess the impact of the universal definition of MI on mortality at 10 years.

Methods and results We studied 676 consecutive patients (Pts) admitted to our intensive cardiac care unit for acute coronary syndrome (ACS) between January 1999 and December 2000. We calculated the relative risk of the total death at 10 years adjusted with the Cox proportional hazards model, between the presence and absence of MI following different definitions: (1), typical symptoms and persistent ST-segment elevation or left bundle branch block (ST-segment elevation definition); (2), typical symptoms and CK-MB activity rise and/or fall >ULN (old definition); and (3), typical symptoms and cTn I rise and/or fall >99th percentile (universal definition). The total mortality at 10 years was 23.8%. The proportion of Pts with AMI was 33.6% for ST-segment elevation definition, 55.8% for old definition, and 70.1% for universal definition. The adjusted hazard ratio of death at 10 years between the presence and absence of AMI was 0.71 (95% confidence interval (CI): 0.46–1.08; P = 0.11) for ST-segment elevation definition, 0.84 (95% CI: 0.55–1.27; P = 0.40) for old definition, and 1.58 (95% CI: 1.07–2.40; P = 0.03) for universal definition. Patients submitted to myocardial revascularization during the initial hospital stay (72%) presented a significantly lower mortality at 10 years, compared with patients not revascularized (adjusted hazard ratio: 0.63, 95% CI: 0.44–0.91; P = 0.014).

Conclusions In a population with the entire spectrum of ACSs, the universal definition of MI increased this diagnosis by one-quarter and was an independent predictor of mortality at 10 years. Furthermore, myocardial revascularization was associated with a significantly lower mortality at 10 years.

  • Acute coronary syndrome
  • Universal definition of myocardial infarction
  • Long-term prognosis
  • Mortality at 10 years

Introduction

Myocardial infarction (MI) is one of the main causes of mortality and morbidity worldwide.1 The occurrence of MI has various major clinical, social, psychological, epidemiological, and research implications.2,3 Clinically, MI has a powerful impact on the quality of life and on patient prognosis.4 The existence of prior MI events has psychological and legal implications. The epidemiological impact of MI derives from the fact that its incidence is used as a health indicator and as an estimate of coronary disease prevalence within a population.5 Finally, MI represents a major clinical event in clinical trials and registries.

In this context, the precise definition of MI assumes a crucial relevance. MI may be defined according to different perspectives and clinical, electrocardiographic, biochemical, imaging, and pathological features.2,3 The development of biochemical markers of high sensitivity and specificity, such as cardiac troponins (cTn) I and T, has represented a remarkable evolution in MI diagnosis and risk stratification610 and it has led to the redefinition of MI by the European Society of Cardiology and the American College of Cardiology in 20002 and, later, to the universal definition of MI by the European Society of Cardiology, the American College of Cardiology Foundation, American Heart Association and the World Heart Federation.3 The criteria for spontaneous MI diagnosis (type 1) came to include cTn elevation, associated with myocardial ischaemia symptoms, above the 99th percentile with a variation coefficient of <10% in an apparently healthy adult population.

The MI definition based on cTn elevation above the 99th percentile has significantly increased MI incidence and it has led to an intense debate on the reclassification of small myocardial necrosis such as MI.1114 Several trials have shown that the new universal definition of MI had implications in the short- and medium-term prognosis11,13,15,16 but there are no evidence regarding the impact on long-term prognosis. The objective of this study is to evaluate the impact of the universal definition of MI on mortality at 10 years.

Methods

Patient selection

All patients consecutively admitted into an intensive cardiac care unit (ICCU) with a final diagnosis of acute coronary syndrome (ACS) between January 1999 and December 2000 were included. ACS was defined as the occurrence of: (i) symptoms suggestive of myocardial ischaemia with duration of at least 10 min at rest within 48 h before admission and (ii) electrocardiogram alterations suggesting myocardial ischaemia, such as ST-segment deviations or negative T-waves in at least two contiguous leads or (iii) elevation of the myocardial necrosis biomarkers.

Clinical assessment and follow-up

Demographic and clinical features were collected during admission and stored in the ICCU database. All electrocardiograms (ECG) were performed according to the ICCU protocol: at admission, in series according to the clinical evolution and with a daily frequency at least. The therapeutic approach and the reference for coronary angiography with possible myocardial revascularization were customized and led by ICCU's assistant physicians and according to the clinical guidelines of the European Society of Cardiology.17 Coronary disease was considered as the presence of stenosis over 50% in epicardial coronary arteries in a coronary angiography. Patients were followed-up by means of routine visits and phone calls whenever clinical file data were deemed insufficient. In 54 patients 10-year vital status was derived from consultation of the National Health System User Card database.

Laboratory assessment

Routine tests (complete blood count, biochemical, and clotting tests) were performed according to the ICCU protocol: at admission, in series according to the clinical evolution and with a daily frequency at least. The measurements of creatine kinase MB (CK-MB) specific cardiac isoform activity and cTn I were performed at admission, every 6 h during the first day and, henceforth, according to the clinical evolution.

Peripheral venous blood samples were collected in dry test tubes and centrifugation was performed in the Laboratory of Biochemistry at 4.000 r.p.m. for 5 min. The measurement of cTn I was performed with the chemiluminescent technique (Access Immunoassay Analyzer, Beckman Coulter, Inc., Fullerton, CA, USA). The lower detection limit for this trial is 0.01 ng/mL. The 99th percentile of cTn I values obtained from a sample of apparently healthy adults is 0.04 ng/mL and the lowest cTn I value for which a maximum variation coefficient of 10% is possible is 0.06 ng/mL. CK-MB analysis was performed at 30°C, following the optimized enzymatic method (multi-analyzer Hitachi 917®), which has an upper limit for normal of 16 UI/L.

The laboratory test results were available ∼1 h after the samples were collected.

MI criteria

Three definitions of MI were used.

ST-segment elevation

The presence of persistent ST-segment elevation in at least two contiguous derivations or any equivalent (left bundle branch block de novo, or presumably de novo, or ST-segment depression in the precordial leads V1–V3) in ECG at admission associated with symptoms suggestive of myocardial ischaemia.

Old definition

The detection of rise and/or fall of CK-MB activity with at least one value above the upper limit of normal (16 U/L) associated with electrocardiographic changes and/or symptoms suggestive of myocardial ischaemia.

Universal definition

The detection of rise and/or fall of cTnI with at least one value above the 99th percentile with a variation coefficient <10% of an apparently healthy adult population (>0.06 ng/mL) associated with electrocardiographic changes and/or symptoms suggestive of myocardial ischaemia.

Study endpoint

The study endpoint was the total mortality at 10 years. We also assessed mortality at 1 and 5 years of follow-up. Vital status at 10 years follow-up was gathered for all patients.

Statistical analysis

The continuous variables with a normal distribution (Shapiro–Wilks test) were expressed as the mean value ± standard deviation and those with a skewed distribution were expressed as the median (interquartile interval). Statistical comparison of baseline features was performed using the χ2 test with the Yates correction for categorical variables, the Student t-test for the continuous variables with a normal distribution and the Mann–Whitney test for the continuous variables with a skewed distribution.

Hazard ratios with 95% confidence intervals (CIs) [hazard ratio (HR) and 95% CI] were calculated for mortality at 1, 5 and at 10 years between the presence and absence of MI according to the three definitions of MI. We performed a sensitivity analysis for the impact of MI definition on mortality at 10 years, through the MI size defined by tertiles of the cTn I peak and by the status of myocardial revascularization performed during the initial admission. Mortality analyses were performed by the Kaplan–Meier technique (log-rank test).

Multivariable analysis was performed with the Cox proportional hazards model to derive potential independent predictors of mortality at 10 years. In these analyses we incorporated the covariates described as baseline demographic and clinical features, the definitions of MI and the status of myocardial revascularization performed during the initial admission. The proportional hazards assumption was checked using the Schoenfeld residual plot vs. time with a smooth curve fit to these residuals. Besides these graphs (one for each covariate) we performed the linear correlation test between time of survival and residual.

For all statistical comparisons, a P value of <0.05 was considered as statistically significant. When appropriate, CIs were calculated using a confidence level of 95%.

Statistical analyses were performed using SPSS version 13.0 (SPSS, Inc., Chicago, IL,USA).

Results

Population characteristics

Six hundred and seventy-six patients with ACS and a mean age of 62 ± 11 years old were included, 80% of which were male. In most patients, the diagnosis consisted of ACS without ST-segment elevation (66.4%).

ST-segment elevation identified 227 patients (33.6%) with MI, the old definition identified 377 patients (55.8%) with MI and the universal definition identified 474 patients (70.1%) with MI.

Patients with the universal definition MI were younger, had a higher prevalence of smoking habits but a lower prevalence of previous MI and percutaneous coronary intervention, higher diastolic blood pressure and heart rate, higher incidence of heart failure signs, and ischaemic electrocardiographic alterations at admission, which resulted in a higher GRACE risk score (Table 1).

View this table:
Table 1

Baseline demographic and clinical features of patients with and without the universal definition myocardial infarction

Baseline featureWith MI, n = 474Without MI, n = 202P
Demographic
 Age (years)61.0 ± 11.763.1 ± 10.40.03
 Male gender (%)81.076.20.18
 Body mass index (kg/m2)26.2 ± 3.926.9 ± 10.60.35
Previous history (%)
 Hypertension57.659.90.61
 Diabetes mellitus21.522.30.84
 Hypercholesterolaemia56.359.40.50
 Smoking32.518.3<0.001
 MI32.743.10.03
 PCI20.537.6<0.001
 CABG14.819.80.11
 Stroke6.35.91
 Peripheral arterial disease8.69.90.43
Admission
 Systolic BP (mmHg)139 ± 29142 ± 270.21
 Diastolic BP (mmHg)82 ± 1779 ± 140.018
 HR (b.p.m.)78 ± 1873 ± 160.001
 Killip class ≥II (%)24.34.0<0.001
 Persistent ST elevation (%)47.90.0<0.001
 ST depression (%)61.043.6<0.001
 Creatinine (mg/dL)1.1 (0.9–1.3)1.1 (0.9–1.3)0.68
 GRACE score140 ± 37115 ± 30<0.001
  • BP, blood pressure; CABG, coronary artery bypass grafting; HR, heart rate; MI, myocardial infarction; PCI, percutaneous coronary intervention.

Patients with the universal definition MI were more frequently treated with thienopyridines, glycoprotein IIb/IIIa inhibitors, and with renin–angiotensin system inhibitors during the initial admission (Table 2). During the initial admission, patients with the universal definition MI presented left ventricular ejection fraction <40% and underwent coronary angiography and percutaneous coronary intervention more frequently but they underwent coronary surgery less frequently (Table 2).

View this table:
Table 2

Pharmacologic treatment, left ventricular ejection fraction, coronary angiography and myocardial revascularization procedures in patients with and without the universal definition myocardial infarction

FeatureWith MI, n = 474Without MI, n = 202P
Pharmacologic treatment (%)
 Aspirin97.595.50.10
 Ticlopidin/clopidogrel64.541.9<0.001
 IIb/IIIa inhibitor54.928.5<0.001
 Heparin96.094.00.31
 Beta blocker88.284.70.31
 ACE inhibitor/ARB82.162.5<0.001
 Statin51.246.00.24
LV ejection fraction <40% (%)a13.36.50.018
Coronary angiography (%)87.680.70.04
 1-vessel disease33.622.40.01
 Multivessel disease62.266.10.40
Myocardial revascularization (%)77.657.9<0.001
 PCI67.138.1<0.001
 CABG11.820.80.01
  • ACE, angiotensin-converting enzyme; ARB, angiotensin-II receptor blocker; CABG, coronary artery bypass grafting; LV, left ventricular; MI, myocardial infarction; PCI, percutaneous coronary intervention.

  • aNot evaluated in 17 patients with MI and 13 patients without MI.

Mortality at follow-up

Mortality results showed the rates of 7.1% at 1 year, 13.6% at 5 years, and 23.8% at 10 years.

Mortality at 1, 5, and 10 years was not significantly different among patients with and without MI following ST-segment elevation criteria and the old definition but it was significantly higher in patients with MI following the universal definition (Table 3 and Figure 1).

View this table:
Table 3

Mortality at 1 year, 5 years and 10 years in patients with and without myocardial infarction using different definitions of myocardial infarction

MI definitionWith MI (%)Without MI (%)HR and 95% CIP
ST elevation
 Death at 1 year6.97.40.94 (0.51–1.74)0.85
 Death at 5 years12.414.60.85 (0.54–1.33)0.48
 Death at 10 years21.724.90.72 (0.50–1.04)0.08
Old definition
 Death at 1 year7.17.11.02 (0.57–1.81)0.95
 Death at 5 years13.513.80.98 (0.65–1.48)0.91
 Death at 10 years23.724.10.91 (0.66–1.25)0.56
Universal definition
 Death at 1 year8.93.92.63 (1.18–5.86)0.01
 Death at 5 years16.19.01.88 (1.13–3.15)0.02
 Death at 10 years26.417.81.50 (1.03–2.20)0.03
  • MI, myocardial infarction.

Figure 1

Cumulative incidence of death at 10 years for patients with and without myocardial infarction using different definitions of myocardial infarction.

In the sensitivity analysis performed in patients with the universal definition MI, the MI size estimated by cTn I peak tertiles did not influence mortality results at 10 years (Table 4). Notwithstanding, performing myocardial revascularization in those patients was associated with a significantly lower rate of mortality at 10 years, compared with patients not revascularized (23.0 vs. 38.1%, P = 0.003).

View this table:
Table 4

Sensitivity analysis for mortality at 10 years in patients with universal definition myocardial infarction according to the myocardial infarction size estimated by cTn I peak and myocardial revascularization procedure at initial admission

CriteriaDeath at 10 years (%)P
MI size
 1st tertile of cTn I peak (0.061–0.772 ng/mL), n = 15830.40.25
 2nd tertile of cTn I peak (0.778–7.55 ng/mL), n = 15822.2
 3rd tertile of cTn I peak (7.57–300.0 ng/mL), n = 15826.8
Myocardial revascularization
 Yes, n = 36823.00.003
 No, n = 10638.1
  • cTn I, cardiac troponin I; MI, myocardial infarction.

Mortality at 10 years adjusted in multivariable analysis

Prior history of hypertension or coronary surgery, a GRACE risk score higher than 148 and the presence of MI following the universal definition were independently associated with a significant increase on mortality at 10 years (Table 5).

View this table:
Table 5

Independent predictors of death at 10 years

Predictive variableHR and 95% CIP
GRACE score >1483.02 (2.12–4.32)<0.001
Hypertension2.28 (1.51–3.43)<0.001
Previous CABG1.68 (1.09–2.57)0.018
MI by universal definition1.58 (1.07–2.40)0.03
Myocardial revascularization during initial hospital stay0.63 (0.44–0.91)0.014
  • CABG, coronary artery bypass grafting; MI, myocardial infarction.

Of all MI criteria analysed (Figure 2), only the universal definition has shown an independent predicting capability for mortality at 10 years (HR: 1.58, 95% CI: 1.07–2.40; P = 0.03).

Figure 2

Adjusted hazard ratios and 95% confidence intervals for death at 10 years in patients with myocardial infarction vs. without myocardial infarction using different definitions of myocardial infarction.

Myocardial revascularization was performed during initial hospital stay in 485 patients (71.7%) and was associated with a significantly lower mortality at 10 years, compared with patients not revascularized (adjusted HR: 0.63, 95% CI: 0.44–0.91, P = 0.014).

Discussion

This is the first study that demonstrates the impact of the universal definition of MI on mortality at 10 years in admitted patients with ACS diagnosis with or without ST-segment elevation. Prognosis implications of MI redefinition and the MI universal definition were already established for hospital mortality13 at 6 months11,16 and at 1 year12 in ACS patients and for major cardiac events at 18 months in patients undergoing percutaneous coronary intervention.15 cTn I and T present a substantially higher sensitivity for myocardial necrosis than CK-MB18,19 along with an almost absolute specificity provided by the fact they were coded by genes specific for the myocardium. The measurement of cTn is a valuable tool for ACS diagnosis in patients with symptoms suggestive of myocardial ischaemia assessed in an emergency unit9 for risk stratification7,8,11 and for ACS patient therapeutic guidance.2022 For all the reasons stated above, cTn has become the standard biochemical method to diagnose MI.2,3

ECG is a first line means of diagnosis to assess patients with symptoms suggestive of myocardial ischaemia.23 In these patients, the presence of persistent ST-segment elevation or any equivalent is generally associated with an acute occlusion of an epicardial coronary artery and the treatment objective consists of obtaining early, complete and sustained reperfusion. Most of these patients present an elevation of myocardial necrosis biomarkers within the first hours of symptom evolution,24 which indicates ST-segment elevation is an indirect MI marker. Notwithstanding the fact that ST-segment elevation in ACS patients is associated with higher in-hospital mortality,25 the differences regarding patients without ST-segment elevation fade during the follow-up.2527 In this study, the definition of MI based on the presence of ST-segment elevation in ECG at admission was not associated with an increase on mortality at 10 years when compared with the absence of MI following that criterion. As referred above, these results can be explained by the higher recurrence of fatal cardiovascular events during the follow-up shown in patients without ST-segment elevation.2527

Prior to MI redefinition in 2000,2 CK-MB cardiac-specific isoform activity elevation above the upper limit of normal was the standard for MI diagnosis.28 However, CK-MB does not detect all degrees of myocardial necrosis, as shown in patients with a sudden death after recurrent events of myocardial ischaemia29,30 or in patients with an unstable angina presenting myocardial necrosis demonstrated by biopsies performed during coronary surgery.31 In our study, the definition of MI based on the detection of rise and/or fall of CK-MB activity above the upper limit of normal was not associated with an increase on mortality at 10 years when compared with absence of MI following that criterion. In this context, the absence of correlation between MI and long-term mortality may be caused by the existence of patients with myocardial necrosis not identified by CK-MB in the group without MI and consequent increase in the risk of fatal cardiac events.7,8,11

In our study, the prognosis implication in long-term mortality of MI universal definition, based on the detection of myocardial necrosis through rise and/or fall of cTnI with at least one value above the 99th percentile of an apparently healthy adult population was not associated with MI size estimated by cTn I peak. These results seem to counter the results of some studies that found a directly proportional relation between the cTn peak and 30-day mortality8 or the combined incidence of death or MI at 6 months,7 but there is no data for long-term mortality as that described in our study. In our population of ACS patients, with and without ST-segment elevation, the universal definition increased the number of patients diagnosed with AMI in ∼25% when compared with the old definition based on the detection of rise and/or fall of CK-MB activity above the upper limit of normal, which is in accordance with the data observed in the literature.11,13,14 The development of cTn assays with increased analytical sensitivity will further increase the proportion of patients with AMI and may improve the ability to identify patients at high risk for major acute coronary events.32

In our population, myocardial revascularization performed during initial admission was associated with a significantly lower mortality at 10 years, after adjustment for confounding factors. In patients with MI following the universal definition treated with percutaneous or surgical myocardial revascularization, the risk of death was 40% lower than in those not revascularized. The association of myocardial revascularization with lower mortality at 10 years observed in our study is in accordance with the results of several other clinical trials comparing invasive strategies with the conservative strategy, which showed that the benefit of systematically performing a coronary angiography depended on the proportion of revascularized patients.33

Among the independent predictors of mortality at 10 years, the GRACE risk score34 was the most powerful, showing that patients with an index >148 presented a risk three times higher than that of the remaining patients. Our study is the first to validate the impact of the GRACE score on mortality at 10 years, confirming the benefit of this index in risk stratification for ACS patients as previously described and recommended.23,35, The universal definition of MI was the sole MI criterion assessed in our study with an independent predicting capability for mortality at 10 years. MI detection through cTn I elevation above the 99th percentile of a healthy population was associated with an increase of ∼60% in long-term mortality, when compared with patients with a true diagnosis of unstable angina.

Limitations

Our study presents some limitations. It is a retrospective study performed in only one centre and with a relatively small population. Studied population is no longer contemporary, as can be verified by the relatively low rate of drug usage established in ACS treatment (thienopyridines and statins). As it was impossible to identify the cardiovascular causes for mortality during follow-up at 10 years, the total mortality was used as the main clinical result. In a recent analysis of the GRACE Registry, cardiovascular causes contributed to 65% of mortality at 5 years but the relative results of total and cardiovascular mortality were similar.27 Regarding the benefit of myocardial revascularization, since it was based on non-randomized data, selection bias is likely to be present.

Conclusions

In a population of patients admitted with the diagnosis of ACS with or without ST-segment elevation, the universal definition of MI showed an independent predicting capability for mortality at 10 years. Furthermore, myocardial revascularization was associated with a significantly lower mortality at 10 years.

Acknowledgements

The statistical analysis support provided by Dr Adriana Belo, biostatistician of the National Cardiology Data Collection Center of the Portuguese Society of Cardiology deserves special recognition.

Conflict of interest: None declared.

References

View Abstract