Response to Letter Regarding Article, “Upsurge in T-Wave Alternans and Nonalternating Repolarization Instability Precedes Spontaneous Initiation of Ventricular Tachyarrhythmias in Humans”
We thank Drs Selvaraj and Chauhan for their interest in our work1 and discussion of the possible impact of white noise and respiration. We do not feel that white noise is a significant problem, although respiration (along with other physiological factors) may play a role, as emphasized in our report.
Impact of Noise
We used the “surrogate” analysis (ie, test of baseline stability) to control for white noise, respiration, movement, and other sources of artifacts. In addition, we removed noisy segments of data and used sectional (T-wave) averaging to reduce the level of white noise. Each data segment was carefully reviewed at high magnification at each processing step. The residual white noise was small (<3 μV), was relatively constant over time, and could not explain the dynamics of T-wave alternans (TWA) or other frequency components. The level of TWA was at least 2 times greater than all sources of noise combined. In addition, the assumption that changes in TWA could be a mere result of noise caused by elevated heart rate and sympathetic activity contradicts the observation that the magnitude of TWA before the arrhythmia was higher than at similar or faster heart rates during arrhythmia-free periods. Therefore, changes in repolarization before the onset of the arrhythmia cannot be explained by an increase in noise at faster heart rates or elevated sympathetic drive.
Impact of Respiration
We explicitly defined the “respiratory” range and emphasized that the greatest increase occurred in that range. Thus, the connection to respiration is clear. However, because the data on respiratory patterns were not available, we avoided speculations about the exact impact of respiration. Furthermore, changes in heart rate in the studied group were relatively small (8%), which does not support the assumption of major respiratory changes.
It is well known that spectral estimates of TWA in real-life recordings are affected by respiration. The magnitude of this effect, which is different for spectral and nonspectral techniques because of the different filtering properties of modified moving average and intrabeat average, needs to be estimated. Therefore, we have provided an explicit estimate of the possible impact of respiratory oscillations on TWA. This estimate shows that the increase in TWA could not be completely explained by this factor. Adding white noise would not change the relationship between spectral components.
Thus, we believe that our presentation of the results is appropriate. Our goal was to describe (not “interpret”) the high-risk periods before the onset of tachyarrhythmias. Although respiration may play some role, the exact role of each contributor requires further study.
Dr Shusterman has ownership interest in PinMed, Inc. A. Goldberg and Dr London report no conflicts of interest.