BOSTON — A novel risk score, comprising measures of wall motion and myocardial perfusion from contrast echocardiography and clinical variables, is a sensitive predictor of 1-year outcome in patients presenting to the emergency department with chest pain prior to obtaining troponin data, reported William Foster, M.D.
In the risk score development model and a subsequent validation model, the tool proved to be more effective for risk stratification than did the use of cardiac troponin measures and clinical variables without the ultrasound data, said Dr. Foster in a poster presentation at the annual meeting of the American Society of Echocardiography.
Dr. Foster and colleagues at the University of Virginia in Charlottesville developed the risk score using clinical and myocardial imaging data from 973 patients presenting to the emergency department (ED) with chest pain that could not easily be attributed to a noncardiac cause and who did not have ST-segment elevation on their admission ECG.
The risk score stratifies the likelihood of developing primary or secondary events within 1 year of chest pain presentation in the ED. Primary events include all cause mortality and myocardial infarction; secondary events include unstable angina, revascularization, and heart failure.
The clinical predictive factors considered in the risk score include age older than 60 years, the presence of three or more coronary disease risk factors, known coronary luminal diameter narrowing of more than 50%, ST-segment deviation on electrocardiogram, two or more angina events in the previous 24 hours, and aspirin use in the previous 7 days. (See box.)
With respect to echocardiographic variables for risk prediction, regional function was characterized as normal or abnormal using a 14-segment model, including 6 segments in each of the basal and midpapillary muscle levels and 2 segments in the apex. Myocardial perfusion was evaluated using the same segmented model and was deemed abnormal if there was no evidence of maximal opacification within a segment by five cardiac cycles. An echocardiographic study was considered abnormal if at least one segment was abnormal for either regional function or myocardial perfusion, Dr. Foster noted.
Each of these predictors is associated with a score between 0 and 100, based on estimates developed using logistic regression models. “The total risk score is the sum of all of these scores,” said Dr. Foster.
Among the 973 patients in the development sample, the model showed “excellent discriminatory capacity,” with an 86% probability of correct prediction, said Dr. Foster. Approximately 60% of those with total risk scores of 200 or higher, and 30% of those with scores of 150–199, experienced a primary or secondary cardiac event at 1 year. About 17% of patients with scores of 100–149, 7% of those with scores of 50–99, and 4% of those with scores of 0–49 had events within 1 year.
To validate the sensitivity of the scores as potential prognostic indicators, the investigators applied the risk score model prospectively in 232 patients who were followed for up to 1 year.
“We saw the same pattern in the validation sample, with a similar prognostic discriminatory capacity between the validation and development samples,” said Dr. Foster. In fact, the discriminatory capacity of the risk score “was greater than clinical variables plus serum cardiac troponin,” he said.
The ability to formulate prognoses prior to obtaining troponin data could streamline the management of chest pain patients in the ED, Dr. Foster concluded.
In this contrast-enhanced image from the apical four-chamber view, arrows show a resting perfusion defect in the mid- and distal septum of a patient with chest pain. Courtesy Dr. Kevin S. Wei