Altered circadian rhythmicity of the QT interval predicts mortality in a large real-world academic hospital population.

Objective And Rationale: Small studies have shown that the QT interval follows a circadian rhythm. This finding has never been confirmed in a large real-world hospital population and the clinical meaning of disrupted rhythmicity remains unknown.

Methods: In this cohort study, all consecutive adult patients with at least one 12-lead ECG acquired between 1991 and 2021 were considered. Sinus rhythm ECGs without QRS conduction or ST-segment abnormalities obtained at the wards or outpatient clinic were included. The QT interval was corrected for age, sex and ventricular rate in a personalized manner. Subsequently, the added value of a 24-h sinusoid of time-of-day was evaluated. An individual 24-h QT interval amplitude was obtained from the model in a subset with patients that had at least 3 ECGs of which one during the night before their last ECG. The association of this individual QT interval with all-cause mortality was assessed using a left-truncated Cox regression model.

Results: The baseline QT correction model was fitted using 237,555 ECGs of 100,644 patients. The personalized corrected QT interval had no relationship with ventricular rate ( = -0.008). Adding the 24-h sinusoidal to the baseline model resulted in a significantly better fit (p < 0.0001). The mean circadian variation of the QT interval was 15 ms, with the maximum QT duration around midnight and an effect that is largest in young female patients. A non-linear relationship between peak-to-trough amplitude in QT interval rhythmicity and all-cause mortality was found, with both lower and higher values associated with increased risk.

Conclusions: Using heterogeneous, real-world hospital data of more than 100,000 patients, circadian rhythmicity proved to be an independent determinant of the QT interval. Both increased and diminished QT rhythmicity was shown to be a predictor of all-cause mortality. QT interval should be corrected for the time-of-day and altered circadian rhythmicity should trigger awareness of increased mortality risk (https://qt.ecgx.ai).