Apolipoproteins were measured in a prospective blinded fashion in blood specimens from patients with chest pain in the emergency department. A definitive diagnosis for the chest pain (non-cardiac-related in 32% and angina or myocardial infarction in 68%) was available in 136 of the 162 patients originally enrolled in the study. Logistic regression and multivariate analysis failed to show any usefulness of apolipoprotein determinations in distinguishing patients with cardiac ischemia from those without it. The clinician's initial impression of the chest pain, the electrocardiogram, a history of previous angina, myocardial infarction, or peripheral atherosclerosis, and male sex were strongly associated with the final diagnosis. We conclude that, although apolipoprotein analysis has proved useful in epidemiologic studies, the most reliable indicators of ischemic pain remain the medical history, the electrocardiogram, and the clinician's overall initial impression.