Can Peak-Picked ECG Be Used for Heart Rate Variability?

Use of peak-picked ECG is generally not recommended for heart rate variability (HRV) calculations because the apparent R-wave peak moves around within the QRS due to the peak-picking operation. This study tests an algorithm for recovering high resolution RR intervals from peak-picked ECG. Two databases were used for testing, simulated ECG and Holter ECG from end stage renal disease (ESRD) patients on and off hemodialysis. Twenty minute samples of single lead ECG (n=1000) were generated by the Physionet HRV ECG simulator. Each ECG record had a random combination of heart rate standard deviation and additive noise. 12-lead, 48 hour Holter ECG from 51 ESRD patients was split into 20 minute segments for HRV analysis. ECG was decimated to 250sps from 1000sps and then peak-picked to the final sample rate of 125sps. High resolution RR interval recovery was based on up-sampling and template matching to align beats. HRV standard deviation of RR intervals (SDNN) was calculated for each 20 min. segment. Bland-Altman analysis was used to assess bias of SDNN error across the SDNN range. The bias of SDNN error at low SDNN values was pronounced for RR intervals defined by R-wave peaks. RR correction by template matching reduced (1) the impact of noise and (2) the SDNN error at low levels of SDNN.