Sleep duration and architecture during ASV for central sleep apnoea in systolic heart failure

Adaptive servoventilation (ASV) effectively treats nocturnal respiratory events in patients with heart failure and reduced ejection fraction (HFrEF) and central sleep apnoea (CSA), but increased mortality has been reported. This study investigated changes in sleep architecture during ASV treatment in HFrEF patients.

A retrospective analysis of polysomnographic datasets for 30 ASV-treated patients with stable HFrEF and moderate-to-severe CSA was performed, including blinded analyses of total sleep time (TST), and percentage of REM and non-REM sleep (stages N1-N3).

Follow-up was 109 ± 32 days; mean device usage was 6.0 ± 1.1 h/day. During ASV there was reduction of N1 (34 ± 20%/TST to 13 ± 5%/TST, p < 0.001) and N3 sleep (4 ± 6%/TST to 1 ± 4%/TST, p = 0.020), and increase of N2 (44 ± 14%/TST to 62 ± 7%/TST, p < 0.001) and REM-sleep (18 ± 8%/TST to 24 ± 6%/TST, p = 0.002).

Disturbances of sympatho-vagal balance during ASV might help explain increased mortality during ASV. Since sympathetic tone is highest in REM-sleep and vagal predominance occurs during N3 sleep, these findings generate new hypotheses for the increased mortality seen in SERVE-HF.