On-line processing of orthogonal exercise electrocardiograms

Abstract

A special program has been developed for on-line processing of orthogonal electrocardiograms during exercise with a small computer system. Special measures were taken to obtain high measurement accuracy and to prevent errors due to baseline drift, skeletal muscle noise or premature beats. The program was tested on exercise electrocardiograms from 47 normal men and 50 patients with coronary artery disease. Single representative complexes were obtained by averaging of beats selected with the aid of the spatial velocity time function of the QRScomplex and the STsegment. The method used is comparable to, but faster than, cross correlation of the spatial velocity time functions. The onset and end of the QRS complex during exercise were determined using templates derived from the first record of the particular subject. This algorithm resulted in a lower measurement error than other methods for waveform analysis.Depression of the ST segment of the electrocardiogram (ECG) recorded during exercise in subjects with a normal ECG at rest is considered to be a specific in,dication for the presence of myocardial &hernia (1, 2, 3). It also has a prognostic value for the development of ischemic heart disease in otherwise asymptomatic subjects (4,5). However, the interobserver variation in the visual interpretation of exercise ECG's is large (6). This is caused in part by the amount of noise present in the signal and in part by the different criteria employed for classification. With digital computer techniques the signal to noise ratio of the ECG and thus the measurement accuracy can be improved. Also multivariate analysis of ECG measurements can now be performed. It is therefore to be expected that computer assisted interpretation will improve the diagnostic value of exercise ECG's (7).Averaging of a number of heart beats is often employed to reduce the amount of noise in the ECG caused by skeletal muscle contractions during exercise. However, averaging may cause considerable errors particularly when the time alignment of the beats is not accurate, or when premature beats are present during exercise. Also QRS complexes may be detected erroneously when sudden baseline shifts take place. Special techniques should therefore be employed to avoid these sources of artifacts.In this study a new method is presented for the detection of QRS complexes and for averaging of the exercise ECG. It is based on selection of those QRS complexes and ST segments which have a similar waveform. In addition an improved method

The Effects of Drugs on the Exercise Electrocardiogram

“…The Frank lead ECG was recorded continuously, and analysed each minute with the aid of a computer system [2], Multiple measurements from three leads were obtained. In this report only ST amplitudes, measured 60 ms after the end of QRS will be presented [3], The ST amplitudes were plotted against heart rate in order to account for the ECG changes that occur in normal subjects which have been shown to be related to heart rate [4], The same plots were made with the rate-pressure product instead of heart rate; however, this did not change the results.…”

A new semiautomated algorithm to quantify holter-detected myocardial ischemia: Preliminary experience in the trimetazidine european multicenter trial (TEMS)

“…The systematic difference in the severity of the STsegment depression measured 0.08 seconds after the J point noted between the Avionics Holter data and our computerized exercise data remains largely unexplained. It is probably due to a systematic error in the detection of the J point, which is determined on a single lead, and probably sooner, with the Holter system, while it is determined with a three-lead, spatial velocity curve in our computerized system for the analysis of the exercise ECG [15,16]. As the difference is systematic and reproducible, it should not interfere with the usefulness of the new algorithm in clinical research.…”

“…In order to test the accuracy of the whole Holter system, calibration signals and two leads (X and Y of the Frank system) were simultaneously recorded during a maximal bicycle exercise test through the same electrode system on a Holter tape and on our computerized system for the analysis and interpretation of the exertional ECG [6,15,16]; this was done in 20 coronary patients, yielding 47 simultaneous measurements for comparison.…”

“…The Holter system was tested with an identical signal recorded on a previously designed and described system for the analysis of the exercise-induced STsegment shifts [6,15,16].…”

Active and passive coping under different degrees of stress; Effects on urinary and plasma catecholamines and ECG T-wave

“…A detailed description of the computerized ECG analysis can be found elsewhere (Simoons, Boom and Smallenburg, 1975). For each sample period a single representative beat with a low noise level was obtained by averaging all detected beats, except those with an aberrant waveform.…”