Device-detected sleep-disordered breathing predicts implantable defibrillator therapy in patients with heart failure

Background: Sleep-disordered breathing is highly prevalent in heart failure (HF) and it has been suggested as a risk factor for malignant ventricular arrhythmias. The Respiratory Disturbance Index (RDI) algorithm computed by select implantable cardioverter defibrillators (ICDs) can identify severe sleep apnea (SA). Purpose: In the present analysis we evaluated the association between ICD-detected SA and the incidence of appropriate ICD therapy in patients with HF. Methods: We enrolled 411 HF patients (age 69±10years, 77% male, ejection fraction 32±8%), implanted with an ICD endowed with an algorithm (ApneaScan, Boston Scientific) that calculates the RDI each night. In this analysis the weekly mean RDI value was considered. The endpoint was the first appropriate ICD shock. The median follow-up was 26 months [25th–75th percentile: 16-35]. Results: During follow-up, one or more ICD shocks were documented in 58 (14%) patients. Patients with shocks were younger (66±13years versus 70±10years, p=0.038), and more frequently implanted for secondary prevention (21% versus 10%, p=0.026). The maximum RDI value calculated during the entire follow-up period did not differ between patients with and without shocks (55±15episodes/h versus 54±14episodes/h, p=0.539). However, the ICD-detected RDI showed a considerable variability during follow-up. The overall median of the weekly RDI was 33episodes/h [25th–75th percentile: 24-45]. Using a time-dependent Cox regression model, the continuously measured weekly mean RDI≥45episodes/h was independently associated with shock occurrence (HR:4.63, 95% CI:2.54-8.43, p<0.001), after correction for baseline confounders (age, secondary prevention). Conclusions: In HF patients, patients were more likely to receive appropriate ICD shocks during periods of time when they exhibited more sleep-disordered breathing.