The diagnostic performance of conventional analysis of the exercise electrocardiographic (ECG) test for the detection of coronary artery disease (CAD) is limited to 70-75%. The exercise ECG test is the most widely used non-invasive method for assessing CAD and even a small improvement in the diagnostic performance is important. The objective of this thesis was to investigate the diagnostic performance of the exercise ECG test for diagnosis of CAD and the prognostic performance of the test for predicting acute myocardial infarction (AMI), in both men and women.
In the diagnostic part of the present study, aimed at finding the optimal diagnostic information during the exercise and recovery phase, 1,877 electrocardiograms selected from 8,322 consecutive patients undergoing a routine exercise test on bicycle ergometer were evaluated. CAD was verified angiographically in 669 patients and excluded in 1,208 patients, by angiography (119 patients), by myocardial scintigraphy (250 patients), or on clinical grounds (839 patients).
The diagnostic performance of ST-segment amplitude, and the difference between heart rate (HR) matched recovery and exercise ST-segment amplitudes (ST/HR difference), were evaluated by constructing receiver operating characteristic (ROC) curves for each sample point every 12th second during 10 minutes of recovery as well as the last 4 minutes of exercise for the ST-segment amplitude. ST-segment amplitude performed better after exercise than during exercise and best within the first 2 minutes of recovery. Its diagnostic ability did not differ from the ST/HR difference. Both methods performed better in men than in women and the diagnostic information appeared mainly in leads I, -aVR, II, V4, V5 and V6.
The effect of different sampling rates (1, 2 and 5 samples/min) on the CAD discrimination ability of ST/HR loop area was evaluated by using ROC curves. A use of sampling frequency below 2 samples/min resulted in a significantly diminished diagnostic performance.
ST-segment and ST/HR variables were compared during exercise and during early post-exercise recovery in terms of diagnostic discrimination capacity and optimal partition values. Discriminating capacity of the methods was evaluated by the use of ROC areas. Partition values, providing the combination of the best sensitivity and specificity, were established. The results demonstrate that analysis of post-exercise ST/HR hysteresis offers the most accurate and gender indifferent identification of patients with CAD.
The prognostic part of the study included 8,317 of the 8,322 patients and covered a mean follow-up period of 9.5 years. This part of the study was used to evaluate the long-term prognostic value of ST/HR hysteresis in predicting AMI and all-cause mortality in men and women. ST/HR hysteresis appears to improve the prognostic ability of an exercise ECG test for AMI and all-cause mortality in a long term perspective compared to conventional ST-segment and ST/HR indicators in both genders and clearly more markedly in women.
Stockholm: KTH , 2009. , 65 p.