Association between implantable defibrillator-detected sleep apnea and atrial fibrillation: The DASAP-HF study.

Abstract

The Respiratory Disturbance Index (RDI) computed by an implantable cardioverter defibrillator (ICD) algorithm accurately identifies severe sleep apnea (SA). In the present analysis, we tested the hypothesis that RDI could also predict atrial fibrillation (AF) burden. Patients with ejection fraction ≤35% implanted with an ICD were enrolled and followed up for 24 months. One month after implantation, patients underwent a polysomnographic study. The weekly mean RDI value was considered, as calculated during the entire follow-up period and over a 1-week period preceding the sleep study. The endpoints were as follows: daily AF burden of ≥5 min, ≥6 h, ≥23 h. Here, 164 patients had usable RDI values during the entire follow-up period. Severe SA (RDI ≥ 30 episodes/h) was diagnosed in 92 (56%) patients at the time of the sleep study. During follow-up, AF burden ≥ 5 min/day was documented in 70 (43%), ≥6 h/day in 48 (29%), and ≥23 h/day in 33 (20%) patients. Device-detected RDI ≥ 30 episodes/h at the time of the polygraphy, as well as the polygraphy-measured apnea hypopnea index ≥ 30 episodes/h, were not associated with the occurrence of the endpoints, using a Cox regression model. However, using a time-dependent model, continuously measured weekly mean RDI ≥ 30 episodes/h was independently associated with AF burden ≥ 5 min/day (hazard ratio [HR]: 2.13, 95% confidence interval [CI]: 1.24-3.65, p = .006), ≥6 h/day (HR: 2.75, 95% CI: 1.37-5.49, p = .004), and ≥23 h/day (HR: 2.26, 95% CI: 1.05-4.86, p = .037). In heart failure patients, ICD-diagnosed severe SA on follow-up data review identifies patients who are from two- to three-fold more likely to experience an AF episode, according to various thresholds of daily AF burden.