Value and limitations of current methods of evaluating patients presenting to the emergency room with cardiac-related symptoms for determining long-term prognosis.

The goal of this study was to determine the value and limitations of the current approach for evaluating patients in the emergency room (ER) with cardiac-related symptoms in terms of predicting long-term outcome. Accordingly, 274 consecutive prospectively identified patients presenting to the ER with such symptoms were evaluated, and follow-up was obtained at 20 +/- 9 months in 265 of them (97%). Adverse cardiovascular events were defined as: nonfatal myocardial infarction, death, cerebrovascular accident with neurologic deficit, life-threatening arrhythmia and cardiac surgery. Eighty-three patients (31%) had a cardiovascular event during follow-up; 42 occurred within 48 hours of ER presentation, whereas 41 occurred in the ensuing months. Findings on physical examination and electrocardiogram provided additional prognostic information, compared with that of history alone, when added sequentially into a Cox model. However, by discriminant function analysis, only 63% of actual events were correctly predicted by the model. Events occurring after 48 hours of ER presentation were correctly predicted only 50% of the time compared with those occurring within 48 hours of ER presentation, which were correctly predicted 75% of the time (p = 0.04). It is concluded that patients presenting to the ER with cardiac-related symptoms are at high risk for adverse cardiovascular events. The likelihood of an event occurring after 48 hours of presentation is as high as one occurring within 48 hours. Current methods of evaluating such patients have limited prognostic value, particularly for those at long-term risk for events.