Aims
Sleep-disordered breathing (SDB) is common in heart failure (HF) and frequently undiagnosed. The ApneaScanTM algorithm, available on certain ICD and CRT devices, uses changes in transthoracic impedance with breathing to quantify SDB. This research tests 3 hypotheses:
1) The ApneaScanTM algorithm can accurately detect moderate-to-severe SDB in patients with HF
2) There is minimal night-to-night variability in the ApneaScanTM-determined severity of SDB
3) Those with moderate-to-severe SDB, assessed by ApneaScanTM, have a higher rate of adverse cardiovascular events than those without.
Methods
Patients with EF≤40% and ICD or CRT devices incorporating ApneaScanTM were recruited. For hypothesis 1, 54 subjects underwent a successful sleep polygraphy study and simultaneous download of ApneaScanTM data. 22 subjects (44%) had undiagnosed moderate-to-severe SDB. The area under the ROC curve was 0.84 for the diagnosis of moderate-to-severe SDB. The optimal ApneaScan cut-off was 30.5/hour (sensitivity 95%, specificity 69%, positive predictive value 68%, negative predictive value 95%). For hypothesis 2, ApneaScanTM data over 28- and 92-nights in 35 patients was reviewed. There was minimal variability in SDB and no significant difference between durations. For hypothesis 3, 72 patients were followed up at a median of 532 (IQR 386-736) days.Mean event-free survival was 660±344 days (95% CI 535-785 days) in the insignificant SDB group and 854±413 days (95% CI 730-978 days) in the significant SDB group (p=0.25 by log rank test).
Conclusions
ApneaScanTM, with an optimal cut-off of 30.5 events/hour, is a sensitive means of screening for SDB in patients with HF with a high negative predictive value. Readings above 30.5/hour require further investigation with a sleep study. Night-to-night variability in SDB is minimal and repeat sleep studies should be reserved for those with ‘borderline’ AHI. In this cohort, the presence of SDB was not associated with adverse cardiovascular outcomes. Recruitment is on-going to test this further.