QT Variability and HRV Interactions in ECG: Quantification and Reliability

Abstract

In this paper, a dynamic linear approach was used over QT and RR series measured by an automatic delineator, to explore the interactions between QT interval variability (QTV) and heart rate variability (HRV). A low-order linear autoregressive model allowed to separate and quantify the QTV fractions correlated and not correlated with HRV, estimating their power spectral density measures. Simulated series and artificial ECG signals were used to assess the performance of the methods, considering a respiratory-like electrical axis rotation effect and noise contamination with a signal-to-noise ratio (SNR) from 30 to 10 dB. The errors found in the estimation of the QTV fraction related to HRV showed a nonrelevant performance decrease from automatic delineation. The joint performance of delineation plus variability analysis achieved less than 20% error in over 75% of cases for records presenting SNRs higher than 15 dB and QT standard deviation higher than 10 ms. The methods were also applied to real ECG records from healthy subjects where it was found a relevant QTV fraction not correlated with HRV (over 40% in 19 out of 23 segments analyzed), indicating that an important part of QTV is not linearly driven by HRV and may contain complementary information.

Techniques for Ventricular Repolarization Instability Assessment From the ECG

“…Multiscale analysis based on the dyadic wavelet transform, allowing representation of a signal's temporal features at different resolutions, has proved useful for QRS detection and ECG delineation [41]. Multi-lead delineation, either based on selection rules applied to single-lead delineation results or based on VCG processing, has shown improved accuracy and stability [42].…”

“…Other approaches to assess repolarization variability use parametric modeling [42], [75], [76]. While in [76] Porta et al investigate variability of the RT interval, in [42] Almeida et al explore QTV, and in [75] the variability from the R peak to the T wave end (RTe) is considered. The use of RT instead of QT avoids the need to determine the end of the T wave, which is usually considered to be problematic.…”

Evidence of cardiac electrical remodeling in patients with Huntington disease

“…Specific variability analyses included:

the standard deviation of normal‐to‐normal RR and QT intervals (SDNN_RR and SDNN_QT in all ECG leads, respectively);

several other time and frequency domain indices of RR interval variability, including the very low (0.0–0.04 Hz), low (0.04–0.15 Hz), high (0.15–0.40 Hz), and total (0.0–0.40 Hz) frequency powers of RR interval variability in natural log‐transformed units (ln ms2/Hz) calculated using autoregression (lnAR) and the Lomb periodogram method (lnLo) (Schlegel et al., 2010);

the QT variability index (QTVI) (Atiga et al., 1998), using the means and variances of the RR interval (Piccirillo et al., 2007) rather than those of the heart rate (Berger et al., 1997) in the denominator of the QTVI equation; and

the “unexplained” part of QTV (Solaimanzadeh et al, 2008; Starc & Schlegel, 2008), wherein the QTV signal is decomposed into two parts, one being described by the concomitant RR interval HRV and/or by the concomitant variability of the QRS‐T angle and the other representing the “unexplained” part of QTV. Decomposition is performed according to a model (Solaimanzadeh et al, 2008; Starc & Schlegel, 2008) that takes into account the hysteresis‐like properties of QT interval dynamics (Lang, Flapan, & Nielsen, 2001) as also the fact that while changes in QT intervals are predominantly driven by changes in RR intervals (Almeida et al, 2006), they can also occur in response to changes in QT wavefront direction descriptors, such as in the QRS‐T angle or equivalent (Acar, Yi, Hnatkova, & Malik, 1999; Kors, van Herpen, & Bemmel, 1999). In a specific manner, we determined the “unexplained” part of SDNN_QT (unexplained SDNN_QT) and the corresponding “unexplained” part of QTV (unexplained QTV).

…”

“…Decomposition is performed according to a model (Solaimanzadeh et al, 2008; Starc & Schlegel, 2008) that takes into account the hysteresis‐like properties of QT interval dynamics (Lang, Flapan, & Nielsen, 2001) as also the fact that while changes in QT intervals are predominantly driven by changes in RR intervals (Almeida et al, 2006), they can also occur in response to changes in QT wavefront direction descriptors, such as in the QRS‐T angle or equivalent (Acar, Yi, Hnatkova, & Malik, 1999; Kors, van Herpen, & Bemmel, 1999). In a specific manner, we determined the “unexplained” part of SDNN_QT (unexplained SDNN_QT) and the corresponding “unexplained” part of QTV (unexplained QTV).…”

Cardiac repolarization analysis using the surface electrocardiogram

“…Consequently, the type of environments for analysis is substantially different from that considered in §5a, in which QT interval adaptation was investigated after possibly large HR changes. Identification of model parameters was performed in Almeida et al (2006), from which QTV fractions correlated and uncorrelated with HRV were quantified. On short-term recordings of healthy subjects, it was found that as much as 40 per cent of QTV was not related to HRV.…”