P959Validation of apneas and hypopneas detection by an implantable cardioverter defibrillator impedance-based respiration sensor: AIRLESS study

Abstract

Background: Among the deleterious consequences of sleep apnea (SA), the most alarming are arrhythmias and sudden death. Diagnosis of SA may reinforce efficacy of cardiac implants for arrhythmias management. Our goal was to validate a transthoracic impedance sensor with an associated autoscoring algorithm for SA diagnosis in ICDs. Objective: To compare 3 months after implantation of ICDs or CRT‐Ds (Boston ScientificTM) the apnea‐hypopnea index (AHI) obtained from polysomnography (PSG) (AHIPSG) with AHI obtained from autoscoring algorithms of the ApneaScanTM implantable impedance respiration sensor (AHIAS). Methods: Patients (pts) were included in a prospective study. PSG scoring was done blinded to AHIAS. Severe SA was defined by an AHIPSG >=30 events/h. Results: 23 pts (22 men, 60 +/‐ 15 years; LVEF 30 +/‐ 7%) were included. Clinical characteristics were: coronaropathy 10/23, atrial fibrillation 9/23, NYHA class III‐ IV: 6/23 pts. 9 pts had a CRT‐D, 14 an ICD. A highly significant correlation was found between AHIPSG and AHIAS especially for the most severe OSA (spearman correlation: 0.78, p <0.0001). Intraclass Correlation Coefficient was in the expected range: 0.69, IC 95% (0.41;0.85). The mean bias (Bland‐Altman analysis) was at 6.3 events per hour (23.3 +/‐ 14.6 vs 29.7 +/‐ 13.7 for mean AHIPSG and AHIAS respectively). Severe SA was diagnosed AHIPSG in 35% of the pts. An optimal AHIAS cutoff value at 30 events/h was obtained by a receiver operator characteristic curve analysis, which yielded a sensitivity of 100% and a specificity of 80%. Conclusion: An advanced algorithm for autoscoring of transthoracic impedance included in ICDs is reliable to identify SA and might improve management of pts with ICD. (Figure presented).