Probability stratification and systematic diagnostic approach for chest pain patients in the emergency department

Management of chest pain patients in the emergency department has been a dilemma because of difficulty in identifying those who can be immediately discharged and those who need to be hospitalized. We assessed the efficacy of a probability stratification model and a systematic diagnostic strategy in 1003 consecutive chest pain patients prospectively evaluated and stratified for acute coronary syndromes according to chest pain characteristics and admission electrocardiogram. Patients with no suspicion of acute coronary syndromes (n = 224) were immediately discharged, whereas those with very-high probability (n =119) were admitted to the coronary care unit. Remaining patients were evaluated in a Chest Pain Unit and investigated during a 9-hour period (intermediate-probability, n = 433) (route 2) and a 6-hour period (low-probability, n = 277) (route 3). Sensitivity and negative predictive value of chest pain type for the diagnosis of acute myocardial infarction (94% and 97%, respectively) was much better than the admission electrocardiogram (49% and 86%, respectively) and admission creatine kinase-MB (46% and 86%, respectively). Serial creatine kinase-MB determinations ruled out acute myocardial infarction by the third-hour postadmission in all route 3 patients but only at the ninth-hour in route 2 patients. For patients with no ST-segment elevation, chest pain type was the strongest independent predictor of acute coronary syndromes. It is concluded that chest pain type is the best single diagnostic tool to rule in/out acute coronary syndromes on admission to the emergency department. Patients with suspicious chest pain must have serum creatine kinase-MB measurements up to 9 hours postadmission to rule out acute myocardial infarction.