The false positive stress test: Multivariate analysis of 215 subjects with hemodynamic, angiographic and clinical data

• Published in: The American journal of cardiology Vol. 40; no. 5; pp. 681 - 685
• Main Authors: Ellestad, Myrvin H., Savitz, Saul, Bergdall, David, Teske, James
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 01.01.1977
• Subjects: Age Factors; Blood Pressure; Cardiac Catheterization; Computers; Coronary Angiography; Coronary Disease - diagnosis; Exercise Test; False Positive Reactions; Female; Heart Rate; Humans; Male; Sex Factors; Statistics as Topic
• ISSN: 0002-9149; 1879-1913
• DOI: 10.1016/0002-9149(77)90182-5

Factors causing the false positive stress test and the ability of the computer to improve test classification were studied in 95 patients with a positive stress test and normal coronary angiograms and 125 patients with a true positive stress test. Multivariate analysis revealed that in men the following clinical findings other than S-T depression were useful in correct stress test classification: (1) maximal heart rate, (2) maximal systolic blood pressure, (3) contour of S-T segment, (4) age, (5) history of chest pain, (6) T waves in resting record, (7) chest pain during test, (8) S-T and T changes with hyperventilation, (9) resting electrocardiogram, (10) time of onset of S-T depression, and (11) increase in P wave negativity in lead V 1 with exercise. These variables, presented in order of importance, had a different ranking in women. With this technique 65 percent of false positive tests were reclassified correctly, whereas they were 100 percent in error using exercise electrocardiographic changes alone. The overall error in all categories was reduced in men from 44 to 21 percent and in women from 63 to 27 percent. Only 13 percent of subjects with a false positive test had no hemodynamic abnormality. The data collected from analyzing the abnormalities in ventricular function suggest that there are few truly false positive tests and that computer processing can significantly improve classification according to the presence of coronary artery disease.