Clinical performance of high frequency QRS analysis for detecting ischemia using a limited sampling rate

In recent years High Frequency QRS analysis (HFQRS) which uses information from the depolarization phase of the cardiac cycle has shown superior accuracy in the detection of stress induced ischemia compared to changes in the repolarization phase. HFQRS analysis typically uses the frequency band of 150-250 Hz and therefore the default requirement was for the ECG to be sampled at 1 kHz. In the present study we examined how HFQRS measurements are affected by a lower sampling rate and evaluated the clinical performance of such analysis. A group of 174 stress ECG records, which were originally sampled at 1 kHz, were analyzed twice using the same HFQRS application: once with the original files and once following decimation to simulate records sampled at 500 Hz. 17% of the patients had significant ischemia and the gold standard was determined according to myocardial perfusion imaging (SPECT). Identical sensitivity of 70% was achieved for both analyses and statistically insignificant difference in specificity was found (81% vs 76%, p=0.39 for the 500 Hz compared to the original records, respectively). The results imply that much of the HFQRS essential information for detecting ischemia exists in the limited frequency band and that the diagnostic performance of the HFQRS are not significantly affected as a result of using a sampling frequency of 500 Hz.