Comparison of three methods for beat-to-beat-interval extraction from continuous blood pressure and electrocardiogram with respect to heart rate variability analysis.

In recent years the analysis of heart rate variability (HRV) has become a suitable method for characterizing autonomous cardiovascular regulation. The aim of this study was to investigate the differences in HRV estimated from continuous blood pressure (BP) measurement by different methods in comparison to electrocardiogram (ECG) signals. The beat-to-beat intervals (BBI) were simultaneously extracted from the ECG and blood pressure of 9 cardiac patients (10 min, Colin system, 1000-Hz sampling frequency). For both data types, slope, peak, and correlation detection algorithms were applied. The short-term variability was calculated using concurrent 10-min BP and ECG segments. The root mean square errors in comparison to ECG slope detection were: 1.74 ms for ECG correlation detection; 5.42 ms for ECG peak detection; 5.45 ms for BP slope detection; 5.75 ms for BP correlation detection; and 11.96 ms for BP peak detection. Our results show that the variability obtained with ECG is the most reliable. Moreover, slope detection is superior to peak detection and slightly superior to correlation detection. In particular, for ECG signals with higher frequency characteristics, peak detection often exhibits more artificial variability. Besides measurement noise, respiratory modulation and pulse transit time play an important role in determining BBI. The slope detection method applied to ECG should be preferred, because it is more robust as regards morphological changes in the signals, as well as physiological properties. As the ECG is not recorded in most animal studies, distal pulse wave measurement in combination with correlation or slope detection may be considered an acceptable alternative.