Incorporating respiratory dynamics into context-aware QTc assessment using a wearable sensor

Abstract Background Traditional short epoch ECG recordings are widely used for QT interval assessment but fail to account for physiological fluctuations that occur over the respiratory cycle. Respiratory-driven variation in QTc interval duration may confound clinical interpretation, especially when QTc prolongation is borderline or under scrutiny for safety monitoring. Objective To quantify the impact of respiratory phase and tidal volume on QTc duration in healthy individuals using continuous, high-resolution wearable data, and to demonstrate the utility of a dry-electrode-based Health Patch for more accurate and physiologically-informed QTc assessment. Methods Twenty-one healthy volunteers (8 females, ages 25–58) wore a dry-electrode Health Patch, which continuously recorded ECG, bioimpedance and accelerometer data during seated rest. A 5-minute interval with least motion was identified using accelerometer data. For this interval, QT and T offset points were identified using an adaptive algorithm robust to ECG morphology variation. QT intervals were corrected using Fridericia’s formula. Respiratory phase and normalized tidal volume were derived from filtered and normalized bioimpedance signals, with phase mapped cyclically from −180° (start-inspiration) to +180° (end-expiration) via the Hilbert transform. A generalized additive model (GAM) was used to model QTc as a function of respiratory phase, tidal volume, and their interaction, using cyclic splines for phase and random intercepts for individuals. Results A total of 6,351 QTc intervals were analyzed (mean 412.4 ms, SD 26.2 ms). QTc duration showed a significant cyclical modulation across the respiratory cycle (Fig. 1), with phase-dependent variations reaching up to 16 ms (p < 0.001). A significant interaction between respiratory phase and normalized tidal volume indicated that deeper breaths enhanced, while shallow breathing attenuated, the phase-related changes in QTc (p < 0.001, Fig. 2). The model accounted for approximately 74% of the variance in QTc deviation. Conclusion This study highlights respiratory activity as a source of systematic, physiological variability in QTc duration that is not captured by conventional short-duration ECG recordings. Such variability, while often overlooked, may influence interpretation in situations where QTc values are near clinical thresholds or under scrutiny for safety monitoring. The Health Patch, by enabling continuous, simultaneous acquisition of ECG and respiratory signals through a wearable system, offers a practical means to incorporate this physiological context into QTc assessment. By capturing beat-to-beat dynamics across multiple respiratory cycles without the constraints of wired systems, it supports a more complete understanding of QTc behavior. This added physiological insight may enhance the reliability of QTc evaluation in settings where accuracy is critical like early-phase clinical environments.Fig 1.Marginal QTc by repiratory phaseFig 1.Marginal QTc by repiratory phase