[Prediction of true and false positive responses to exertion by a second exercise under sublingual nitroglycerin....

The authors consider that the diagnostic value of exercise tests could be improved by a multivariate analysis integrating the ergometric data of two exercise tests: a reference test and a test performed under nitroglycerin. 109 patients without previous myocardial infarction who suffered from angina-like chest pain were explored. All had an ischaemic-type ST depression on the CM5 lead during the first exercise test and were tested again after sublingual administration of nitroglycerin. Both tests were continued until the theoretical maximum heart rate was reached or symptoms were no longer bearable. Coronary arteriography showed that 27 patients had normal coronary vessels and 82 had a coronary disease. At univariate analysis, two parameters were significant in predicting coronary disease. These were an in increase, between the two tests, of maximal work performed (from 95 +/- 1578 kpm to 2085 +/- 1662 kpm, p less than 0.001) and of the double product, i.e. maximum heart rate x maximum systolic arterial pressure (from 181 +/- 5289 to 3826 +/- 4245 mmHg.b.min-1, p less than 0.001). Five variables were selected by logistic regression analysis, viz.: change in ST depression between the two tests, double product during the first test, sex, modification of work performed between the two tests and occurrence of chest pain during the first test. By using a threshold analysis (less than 20 p. 100, greater than 80 p. 100), 76 p. 100 of the population was correctly classified; 82.9 p. 100 of coronary disease patients and 53 p. 100 of normal subjects were well classified, whereas 2.7 p. 100 of the population studied (all normal subjects) were erroneously classified.(ABSTRACT TRUNCATED AT 250 WORDS)