Observational studies have reported a marked discrepancy between the risk estimated by scores and the use of an invasive strategy in patients with acute coronary syndromes. The objective is to describe the criteria used to decide an early invasive strategy and to determine the differences between those criteria and the thrombolysis in myocardial infarction risk score (TRS). Patients entered to the Epi-Cardio registry with a diagnosis of non–ST-elevation acute coronary syndrome were analyzed. A logistic regression model including variables associated with an early invasive strategy was developed and validated in 2 consecutive cohorts. The association between the TRS and the clinical decision model with an early invasive strategy was evaluated by receiver operating characteristic (ROC) curves. We included a total of 3,187 patients. In the derivation cohort, variables associated with an early invasive strategy were previous angioplasty (odds ratio [OR] 1.63), hypercholesterolemia (OR 1.36), ST changes (OR 1.49), elevated biomarkers (OR 1.42), catheterization laboratory availability (OR 1.7), recurrent angina (OR 3.45), age (OR 0.98), previous coronary bypass (OR 0.65), previous heart failure (OR 0.40), and heart rate at admission (OR 0.98). The areas under the ROC curves to predict invasive strategy were 0.55 for the TRS and 0.69 for the clinical decision model, p <0.0001. In the validation cohort, ROC areas were 0.58 and 0.70, respectively, p <0.0001. In conclusion, invasive strategy was guided by variables not completely included in risk scores. The clinical, evolutionary, and structural variables included in the model can explain, partially, the discordance existing between risk stratification and medical strategies.
Acute coronary syndromes with non–ST-segment elevation (NSTE-ACS) are the most frequent clinical presentation of acute coronary heart disease, comprising approximately 60% of cases. Risk scores have been developed for the initial risk stratification of these patients. Several randomized studies have shown a trend toward greater clinical benefit when implementing an early invasive strategy in higher risk patients who were stratified using such scores. However, observational studies have reported a marked discrepancy between the risk estimated by various scores and the use of an invasive strategy. Several hypotheses have been put forward to justify such discrepancy between clinical practice and the recommendations made by international guidelines. The objective of our study is to describe the criteria used to decide an early invasive strategy in the real world and to determine the differences between those criteria and the thrombolysis in myocardial infarction risk score (TRS).
Methods
Epi-Cardio is a prospective, multicenter, self-financing registry of cardiovascular intensive care units (intensive care and coronary care units) of Argentina. Participating centers voluntarily joined the registry, when invited to do so at medical congresses and meetings of the cardiology professional societies of our country (Sociedad Argentina de Cardiología and Federación Argentina de Cardiología), or through direct registration through the project’s web page (for further information, see www.epi-cardio.com.ar ).
Our study complies with the principles of the Declaration of Helsinki. A locally appointed ethics committee approved the research protocol.
The study population consisted of all patients admitted to the institutions participating in the Epi-Cardio Registry with a diagnosis of NSTE-ACS from October 2006 to January 2010. Patients admitted in 2006–2007 comprised the derivation cohort used to develop an early invasive strategy model (EISM), and the following patients admitted in 2008 to 2010 comprised the validation cohort.
Patients with a diagnosis of secondary myocardial ischemia (anemia and tachyarrhythmias) were excluded. Patients were stratified as high, moderate, or low risk according to the TRS at admission. The TRS is calculated by adding 7 dichotomous variables (a point for each variable present): age ≥65 years, 3 or more cardiovascular risk factors present, use of aspirin during the 7 days before admission, known coronary stenosis >50%, 2 or more episodes of angina in the last 24 hours, ST-segment changes, and elevated markers of myocardial injury (creatine kinase MB fraction and/or cardiac-specific troponin level). A score of 0 to 2 indicates a low clinical risk, 3 to 4 is moderate risk, and 5 to 7 indicates high risk of myocardial infarction or death.
Patients were classified according to the therapeutic strategy used: early invasive strategy or early conservative strategy. The early invasive strategy group included patients who underwent coronary angiography during the initial 72 hours post admission, without a previous stress test to detect ischemia. The initial conservative strategy group included patients who did not undergo coronary angiography or patients who were catheterized after 72 hours post admission or patients who underwent a stress test to detect ischemia.
Baseline characteristics of patients who were treated with both strategies were compared among the derivation cohort. Continuous variables are expressed as mean and SD in cases with normal distribution and as median and interquartile interval for cases with non-Gaussian distribution. Categorical variables are expressed as percentages. For continuous variables, differences among groups were assessed with the t test or Mann-Whitney U test, according to their distribution. Categorical variables were compared using contingency tables. To assess the relation between the use of the early invasive strategy and the increasing levels of risk, we used the Cochran-Armitage test. To assess the selection criteria for the early invasive strategy, with the derivation cohort, we built the EISM, which consists of a logistic regression model where we incorporated all the variables associated with the use of an early invasive strategy with a p value <0.1 on univariate analysis. Finally, the model included all variables that were significantly associated with the decision of an early invasive strategy on multivariate analysis. The ability of this model to predict the use of an invasive strategy (discrimination) was assessed and compared with the TRS using receiver operating characteristic (ROC) curves. The EISM was tested in the validation cohort. Calibration of the model was assessed with Hosmer-Lemeshow’s test in the derivation and validation cohorts. A p value of <0.05 was considered statistically significant for all 2-tailed comparisons. The statistical analysis was performed with STATA 10.0 (StataCorp LP., College Station, Texas) and Epi Info 3.5.1 software.
Results
A total of 3,187 patients from 43 institutions distributed in 11 argentine provinces (with a total of 623 institutions with acute care unit) were included. Thirty-five centers (81%) have a catheterization laboratory and included 75.7% of the total number of patients assessed. Characteristics of the population are listed in Table 1 .
| Variable | Cohort Derivation (n = 1,396) | Cohort Validation (n = 1,791) | p Value |
|---|---|---|---|
| Age (yrs) | 65 (55–74) | 62 (54–71) | <0.001 |
| Men | 938 (67.2) | 1,264 (70.6) | 0.04 |
| Hypertension ∗ | 972 (69.6) | 1,206 (67.3) | 0.16 |
| Diabetes mellitus | 314 (22.5) | 375 (20.9) | 0.29 |
| Hypercholesterolemia † | 685 (49.1) | 910 (50.8) | 0.33 |
| Active smokers | 374 (26.8) | 482 (26.9) | 0.94 |
| Previous infarction | 337 (24.1) | 373 (20.8) | 0.02 |
| Previous coronary bypass | 144 (10.3) | 161 (9) | 0.21 |
| Previous coronary angioplasty | 264 (18.9) | 325 (18.1) | 0.58 |
| Previous chronic angina | 214 (15.3) | 233 (13) | 0.06 |
| Impaired left ventricular systolic function ‡ | 261 (32.4) | 273 (28.6) | 0.08 |
| ST-segment changes | 402 (28.8) | 456 (25.5) | 0.03 |
| Elevated biomarkers | 429 (30.7) | 737 (41.2) | <0.001 |
| High-risk TIMI risk score | 129 (9.2) | 125 (7) | 0.0001 |
| Moderate-risk TIMI risk score | 740 (53.0) | 861 (48.1) | |
| Low-risk TIMI risk score | 527 (37.8) | 805 (44.9) | |
| Systolic blood pressure at admission (mm Hg) | 135 (120–150) | 135 (120–150) | 0.87 |
| Hart rate at admission (beats/min) | 71 (60–80) | 70 (60–80) | 0.82 |
| Major co-morbidities | 277 (19.8) | 372 (20.8) | 0.52 |
∗ Blood pressure ≥140/90 mm Hg or previous pharmacologic treatment, according to the local guidelines.
† Total cholesterol ≥240 mg/dl or previous pharmacologic treatment, according to the local guidelines.
‡ Among 806 patients in the derivation cohort and 953 patients in the validation cohort, with echocardiographic assessment of left ventricular systolic function.
According to the TRS, among the 1,396 patients in the derivation cohort, 129 patients (9.2%) were considered to be high risk, 740 (53%) were at moderate risk, and 527 (37.8%) were at low risk. Among the 1,791 patients in the validation cohort, 125 (7%), 861 (48%), and 805 (45%) patients were at high, moderate, and low risks, respectively.
The invasive strategy was used in 1,676 patients (52.6%), 742 patients (53.2%) in the derivation cohort, and 934 patients (52.1%) in the validation cohort, p = 0.57. This strategy was indicated more often in centers with a Catheterization Lab available (57.9% vs 42.8%, p <0.00001, in the derivation cohort and 55.8% vs 36%, p <0.00001, in the validation cohort, respectively). Use of the invasive strategy according to the TRS was 57.4%, 55.5%, and 48.8% (p for trend <0.05) in the high, moderate, and low risks, respectively, in the derivation cohort, and 63.2%, 57%, and 45.2% (p for trend <0.00001), respectively, in the validation cohort.
The association between different variables and the indication of an early invasive strategy assessed with univariate analysis in the derivation cohort is listed in Table 2 . Variables associated with the early invasive strategy in the logistic regression analysis were listed in Table 3 . All these variables were included in the EISM. The EISM exhibited a greater association with the invasive strategy than the TRS. In the derivation cohort, the area under the TRS ROC curve to predict an early invasive strategy was 0.55 and that of the EISM was 0.69 (p <0.0001). Among the validation cohort, values were 0.58 for the TRS and 0.70 for the EISM (p <0.0001; Figure 1 ). Calibration of the model was adequate, as with Hosmer-Lemeshow’s test, p values were 0.66 for the derivation cohort and 0.44 for the validation cohort. Discrimination of the model was moderate for both cohorts. The model was satisfactorily validated; areas under the ROC curve were similar for both cohorts. The EISM showed an association with the early invasive strategy that was similar when applied only to centers with availability of a Catheterization Laboratory (area under the ROC curve: 0.69 for both cohorts).
| Variable | Early Invasive Strategy (n = 742) | Early Conservative Strategy (n = 654) | p Value | OR (95% CI) |
|---|---|---|---|---|
| Age (yrs) | 63 (55–72) | 66 (56–77) | <0.0001 | |
| Men | 68.6 | 65.6 | 0.23 | 1.15 (0.92–1.43) |
| Hypertension | 68.7 | 70.6 | 0.43 | 0.91 (0.73–1.15) |
| Diabetes mellitus | 22.5 | 22.5 | 0.99 | 1.00 (0.78–1.29) |
| Hypercholesterolemia | 53.2 | 44.3 | 0.0009 | 1.43 (1.16–1.76) |
| Active smokers | 30.1 | 23.1 | 0.003 | 1.43 (1.13–1.82) |
| Three or more risk factors | 19.7 | 18 | 0.44 | 1.11 (0.85–1.46) |
| Previous infarction | 24.1 | 24.2 | 0.99 | 0.99 (0.78–1.28) |
| Previous coronary bypass | 8.8 | 12.1 | 0.04 | 0.70 (0.49–0.98) |
| Previous coronary angioplasty | 21.4 | 16.1 | 0.01 | 1.43 (1.09–1.88) |
| Chronic stable angina | 16.2 | 14.4 | 0.35 | 1.15 (0.86–1.54) |
| Impaired left ventricular systolic function ∗ | 34.1 | 30.1 | 0.23 | 1.20 (0.89–1.54) |
| ST-segment changes | 33 | 24 | <0.001 | 1.56 (1.23–1.98) |
| Elevated biomarkers | 34.4 | 26.6 | <0.01 | 1.44 (1.15–1.82) |
| High-risk TIMI risk score | 57.4 | 42.6 | 0.035 | |
| Moderate-risk TIMI risk score | 55.5 | 44.5 | ||
| Low-risk TIMI risk score | 48.8 | 51.2 | ||
| Systolic blood pressure at admission (mm Hg) | 133 (120–150) | 135 (120–154) | 0.74 | |
| Heart rate at admission (beats/min) | 70 (60–80) | 75 (60–80) | 0.0003 | |
| Major co-morbidities | 147 (19.8) | 130 (19.9) | 0.97 | 1.00 (0.85–1.23) |
| Coronary angiography | 742 (100) | 25 (3.8) | ||
| Mortality | 1.6 | 2.4 | 0.27 | 0.65 (0.30–1.40) |
∗ Among 806 patients in the derivation cohort, with echocardiographic assessment of left ventricular systolic function.
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
