ICD-Detected Sleep-Disordered Breathing and Shocks in Heart Failure Patients

Quick Takes

  • ICDs with algorithms able to detect sleep-disordered breathing can continuously quantify the severity of the disease.
  • In patients with HFrEF, measures of ICD-detected sleep-disordered breathing are associated with ICD shocks.

Study Questions:

What is the association between implantable cardioverter-defibrillator (ICD)–detected sleep-disordered breathing (SDB) and incidence of ICD shocks in patients with heart failure (HF)?


This was a prospective, observational, multicenter study (RHYTHM DETECT) of consecutive patients with HF with reduced ejection fraction (HFrEF) that underwent ICD alone or cardiac resynchronization therapy defibrillator (CRT-D) device implantation. Patients were included if their device had the SDB diagnostic algorithm to detect obstructive and central sleep apnea. Patients also needed to be enrolled in the device’s remote monitoring platform. The SDB diagnostic algorithm utilized continuously measured thoracic impedance to determine apnea and hypopnea episodes at night to calculate a Respiratory Disturbance Index (RDI) [episodes/hour]. Maximum and weekly mean RDI values were reported. The primary endpoint was first appropriate ICD shock. The secondary endpoint was first appropriate ICD therapy (anti-tachycardia pacing [ATP] or shock).


A total of 411 patients from 27 centers were enrolled between December 2017 and June 2021. Median follow-up was 26 months (interquartile range [IQR] 16-35). At baseline, the patients were an average age of 69 years (standard deviation [SD] ± 10), 77% male, average body mass index 26 kg/m2 (SD ± 4), 47% with ischemic heart disease, New York Heart Association class II (61%) or III (33%), average EF 32% (SD ± 8), 12% with secondary prevention ICD indication, and 72% with CRT-D. No patients were receiving continuous positive airway pressure therapy during the observation period. Median weekly RDI was 33 episodes/hour (IQR 24-45). Average maximum RDI over the entire study period was 54 episodes/hours (SD ± 14).

For the primary endpoint, 14% (58/411) of patients received an ICD shock. The median weekly RDI during weeks with an ICD shock compared to no shock was significantly higher (47 vs. 33 episodes/hour, p = 0.026). However, the maximum RDI over the entire study period for patients receiving ICD shock when compared to those who did not receive a shock was not significantly different (55 vs. 54 episodes/hour, p = 0.539). For patients with a weekly mean RDI >45 episodes/hours compared to those at ≤45 episodes/hour, the rates of shocks over time were higher (0.20 vs. 0.04 shocks/year) with an incidence rate ratio of 4.53 (95% confidence interval, 2.43-8.62; p < 0.001). On univariate analysis, younger age and ICD for secondary prevention were also associated with ICD shock. No significant changes were noted after adjusting for these factors on multivariate analysis.

For the secondary endpoint, 24% (100/411) of patients received an ICD therapy (ATP or shock). Associations between the secondary endpoint and median weekly RDI, maximum RDI over the study period, and weekly mean RDI >45 episodes/hours compared to ≤45 episodes/hour were similar to the primary endpoint.


The investigators concluded that in patients with HFrEF, measures of device-detected SDB were associated with ICD shock.


SDB, including both obstructive and central sleep apnea, is common in patients with HF and associated with poor prognosis. SDB is also noted to be a risk factor for ventricular arrhythmias. This study interestingly notes that an ICD-based method for quantifying SDB over a long period of time demonstrates great variability in the weekly severity of SDB for an individual. For predicting ICD therapies, this viability may in part explain why the single maximum RDI for a patient was not as useful compared to other RDI measures. Keeping in mind the limitations of this study (lack of polysomnography data and formal SDB diagnosis, unclear HF status at time of ICD therapy, variable ventricular arrhythmia detection and therapy settings), future research will need to conducted to better understand the role of serial monitoring of device-detected SDB in the management of sleep apnea, HF, and potential ventricular arrhythmias.

Clinical Topics: Arrhythmias and Clinical EP, Geriatric Cardiology, Heart Failure and Cardiomyopathies, Prevention, Implantable Devices, SCD/Ventricular Arrhythmias, Atrial Fibrillation/Supraventricular Arrhythmias, Acute Heart Failure, Sleep Apnea

Keywords: Apnea, Arrhythmias, Cardiac, Body Mass Index, Cardiac Resynchronization Therapy, Continuous Positive Airway Pressure, Defibrillators, Implantable, Electric Impedance, Geriatrics, Heart Failure, Myocardial Ischemia, Pacemaker, Artificial, Secondary Prevention, Shock, Sleep Apnea, Central, Sleep Apnea, Obstructive, Sleep Apnea Syndromes, Stroke Volume, Tachycardia

< Back to Listings