Improvements in exercise electrocardiography.

In 1908, Einthovent published the first example of normal electrocardiographic responses to exercise. In 1928, Feil and Siegel2 observed abnormal downsloping ST depression and T wave inversion in three of four patients experiencing anginal attacks (presumably after climbing stairs to the phy-sician's office; H. Hellerstein, personal communication). In 1932, Goldhammer and Scherf3 associated this response with coronary insufficiency, and in 1941, Master and Jaffee4 added postexertional elec-trocardiographic (ECG) monitoring to the previously established two-step test. In 1965, Blomqvist5 made important observations on changes in computerized Frank lead ECG recordings, and in 1966, Bruce See p 245 and colleagues6 evaluated variance of computer-averaged and computer-analyzed components of the ECG. Hornsten and Bruce,7 in 1969, compared computerized bipolar precordial and Frank lead responses in healthy young and middle-aged men with those of patients who had coronary heart disease. Also in 1969, Bruce and McDonough8 reported quantitative differences in computerized ST responses during exercise and hysteresis variations in recovery and plotted ST displacement versus heart rate as stress-strain diagrams. Since then, innumerable clinical and research studies have described the value and limitations of exercise electrocardiography, one example being published in 1984.9 Common problems regarding the value of exercise electrocardiography include "false-positive" responses in healthy persons, especially in women, and "false-negative" responses in patients with coronary heart disease. A new and interesting insight into the diagnostic value of exercise electrocardiography was the prediction of severity of coronary artery disease with the slope of the submaximal ST segment/heart rate relation that Elamin and colleagues10 published in 1980. Despite initial skepticism, subsequent publications from these investigators in Leeds, England, in 1982 and 1984 have strengthened the evidence for diagnostic value and have prompted others to adopt the improved method that incorporates exertional ST responses. Among the latter group are two important contributions. In 1986, Okin, Amiesen, and Kligfieldl1 emphasized the value of modifying the conventional multistage Bruce protocol for tread-mill exercise testing to provide smaller increments in heart rate with increasing workloads. This modification provided more computerized data points (every 2 minutes) to better determine the greatest final changes in the ST-segment/heart rate slope. The mean increment in heart rate by their modified protocol was only 10 beats/min rather than 27 beats by the conventional Bruce protocol, yet overall heart rate and systolic pressure changes were similar. They established both the accuracy of computer measurements of the ST segment at 60 msec after the J point …