Analysis of physical activity and sleep behaviour preceding malignant ventricular arrhythmias treated by implanted cardioverter-defibrillators

Physical activity and sleep behaviors are prognostic for ventricular arrhythmias in ICD patients. However, short-term patterns in these behaviors in the days before ICD therapy remain underexplored. By analyzing accelerometer data in the month before a ventricular arrhythmia, we aim to find actionable predictors of arrhythmias and clinical deterioration in ICD carriers.

This study aims to investigate behavioural patterns immediately preceding malignant ventricular arrhythmias in patients with an ICD by analysing day-to-day behavioural data.

This is a sub-analysis from a multicenter, prospective study, conducted at two European centres. Patients with an ICD were enrolled between May 2021 and September 2022 and wore wrist-based activity monitors for 12 months. Thirteen metrics of daily activity and sleep were derived from tri-axial accelerometer data. Baseline averages were calculated from the first 30 days of monitoring; deviations from these baselines were analyzed over the last 30 days before the outcome of appropriate ICD therapy (shock or anti-tachycardia pacing). A Cox Proportional Hazards model with time-varying variables averaged over each 30-day period was used to identify risk of outcome based on activity and sleep patterns over time. Patients not wearing the wristband 30 days pre-therapy were excluded. The model was adjusted for age, secondary or primary prevention, ischaemic or nonischaemic cardiomyopathy, device type (ICD or CRT-D), and heart failure diagnosis (reduced or preserved ejection fraction).

A total of 277 patients (mean age 63 ± 10 years, 81% male) were included, with 51 (18%) experiencing appropriate ICD therapy during follow-up. Long-term analysis revealed no significant differences in average one-year behavior metrics between patients with and without the outcome. In time-varying analysis, two patterns appeared: higher intensity during the most active six minutes each day (M6) was associated with a lower risk of imminent ICD therapy (HR 0.47, 95% CI 0.24–0.94), while longer exercise duration was linked to higher risk (HR 2.16, 95% CI 1.03–4.59). This inverse association suggests that longer but less intense activity durations can be a sign of disease deterioration. Figure 2 shows four examples of baseline deviations, with no significant differences between baseline and 30 days before ICD therapy.

This sub-analysis of a prospective study demonstrated changes in intensity and duration of activity to be associated with imminent ventricular arrhythmia. The findings suggest that both the intensity and duration of daily activity impact arrhythmia risk differently, confirming the role of physical activity in arrhythmia risk and generating further hypotheses for understanding individualized risk assessment and personalized treatment.

Forest plot of the hazard ratios