Keywords
INTRODUCTION
The Joint European Society of Cardiology/American College of Cardiology Committee's (ESC/ACC) new definition of myocardial infarction (NDMI)1 has important clinical, epidemiologic, legal and scientific implications.2 Moreover, it has been seriously questioned since its publication in 2000.3,4 In Spain, the survey carried out by the Working Group on Ischemic Heart Disease and Coronary Care Units of the Spanish Society of Cardiology showed that 42% of Spanish cardiologists rarely or never applied the NDMI.5 Currently, we do not know how widely the NDMI is used. However, the variety of ways in which the NDMI is interpreted must mean that neither population studies nor healthcare plans reflect the realities of myocardial infarction (MI). This may well lead to incorrect conclusions and misguided patient management strategies.
In this study we describe the extent to which the NDMI is used in daily clinical practice.
PATIENTS AND METHOD
Cross-sectional, observational study of patients discharged from the cardiology department of a university hospital in Murcia, sSoutheastern Spain.
Patients
Between March 1 and August 31 2001, 277 patients, diagnosed on admission with suspected acute coronary syndrome (ACS), were discharged. We reviewed their clinical records and excluded 150 patients who did not satisfy the NDMI (Table 1). The final sample consisted of 127 patients.
Definitions and variables
We divided our sample into two groups: group A patients met both traditional MI diagnostic criteria (Table 2) and the NDMI; group B patients met only NDMI criteria.
We considered increases in the following enzymes as significant when accompanied by a characteristic rise and fall typical of MI: CPK≥300 ng/dL; CPK-MB mass≥5 ng/dL; cardiac troponin T≥0.1 ng/dL.
We included patients in the sample when the diagnosis on admission either used the term MI explicitly or included a term synonymous with myocardial necrosis.
Statistical analysis
Quantitative variables are expressed as the mean and SD. We constructed a logistical regression model to identify variables independently associated with a hospital discharge diagnosis of MI. We analyzed data with SPSS 11.0 for Windows.
RESULTS
Of the 127 patients discharged during the period of our study after MI as defined by the NMDI, 77 also met classical criteria for MI (61%, group A). The remaining 50 patients (39%) presented isolated elevated levels of the new markers, cardiac troponins or MB mass (group B). Baseline characteristics of groups A and B appear in Table 3. Table 4 shows clinical characteristics of MI presentation. Group A showed significantly greater ischemic abnormalities in the ECG, peak CPK and duration of symptoms. Table 5 shows the patients' hospital course and diagnostic and therapeutic tests on admission. We found no differences between diagnostic tests except for a significantly greater use of coronary angiography in Group A. Group A patients showed significantly greater use of reperfusion treatment, greater frequency of percutaneous revascularization, and longer average hospital stay. The diagnosis of MI by imaging techniques was successful significantly more frequently in group A than group B (91% vs 60%; P<.0005). The diagnosis of MI at discharge was recorded for 98 patients (77% of the sample); 74 (96%) of them were from group A and 24 (48%) from group B (Figure 1). Eight patients showed increased enzyme levels after coronary angioplasty but only 3 (37.5%) of them of were diagnosed as having MI after the procedure. All three belonged to group A.
Fig. 1. Distribution of final clinical diagnoses in discharge reports for group A, group B and groups A and B together.
Diagnoses in the 29 patients not classified as suffering from MI appear in Figure 2.
Fig. 2. Final clinical diagnoses in discharge reports of patients not diagnosed as having myocardial infarction even though they met the criteria in the new definition
According to the multivariate analysis, belonging to group A, confirmation of MI by imaging techniques, and acute ischemic abnormalities in the ECG were associated independently with a greater likelihood that the specialist would classify events as MI (Table 6). We developed a model that accounted for 92% (CI 95%, 88%-97%) of the clinical diagnoses of MI (area below the [ROC] curve).
DISCUSSION
Our results show how little the NDMI has influenced the criteria for diagnosis at discharge from the cardiology department of our university hospital. Of 127 patients who satisfied the NDMI, 77% were diagnosed accordingly at discharge versus only 48% of patients added by applying the new definition.
The prevalence of MI among patients discharged after admission for ACS was 35% in our sample. This is 24% less that it would have been (46%) if the NDMI had been rigorously applied. The ESC/ACC committee predicted an increase in the number of infarctions diagnosed by applying NDMI criteria1 and highlighted this in subsequent research.6,7 This confirms the impact of the NDMI on epidemiologic data, distribution of healthcare resources and patient management.
Our data, which reflect the actual level of implementation of the NDMI criteria, support the opinion of Spanish cardiologists expressed in a recent survey.5 In our sample, patients who did not meet classical criteria for the diagnosis of MI had similar clinical antecedents, except for the significantly lower frequency of a history of ischemic heart disease, and fewer clinical symptoms (lower frequency of ECG changes, duration of symptoms, peak enzyme levels, length of hospitalization). In addition, hospital discharge diagnoses may have been more strongly conditioned by the smaller number of cases of infarction that were confirmed with imaging techniques. Essentially, we were dealing with smaller infarctions according to the generally accepted grading system of infarction size. Moreover, these myocardial infarctions were clinically less important in terms of hospital course. The multivariate model we constructed explained 92% of the decisions (ROC). In the model, the presence of ischemic abnormalities in the ECG, including those defined by criteria not contemplated by the ESC/ACC joint committee such as confirmation of MI with imaging techniques, were the main factors that conditioned the final diagnosis of MI in patients with increased enzyme levels but with none of the traditionally recognized symptoms of MI. We found that the presence of new markers of necrosis (cardiac troponins or CK-MB) together with symptoms of myocardial ischemia was not sufficient in the majority of patients for the specialist to establish a final diagnosis of MI at discharge.
Limitations
Our study is based on a single department, so the results may not be representative of other centers in Spain. The lack of prior agreement on the implementation of the NDMI within our hospital, where 5 or more specialists make diagnoses independently, suggests that NDMI use may be similar in other Spanish centers.
Our study covers a period of 6 months. With time, results might change, and use of the NDMI might increase or decrease. We do not believe this invalidates our results; rather, it reinforces one of our conclusions: the variety of ways in which the NDMI is implemented will not only make it more difficult to compare centers, but will also complicate comparisons within a single center at different periods of time.
CONCLUSIONS
Implementation of the NDMI for hospital discharge diagnoses is limited. This definition is used for fewer than 50% of patients who do not meet classical criteria for the diagnosis of MI. Ischemic abnormalities in the ECG and diagnosis with imaging techniques are the principal factors that lead to the presenting event being considered MI in patients who do not meet classical MI criteria.
Part of this report was presented at the XXXVIII Congreso de Enfermedades Cardiovasculares (XXXVIII Conference on Cardiovascular Diseases), Madrid, October 2002.
Correspondence: Dr. R.F. López Palop.
Ricardo Gil, 20, 3° dcha. 30002 Murcia. España.
E-mail: mlopezs@meditex.es