Detection of apnoeas, hypopnoeas and arousals by the AutoSet in the sleep apnoea/hypopnoea syndrome

Limited sleep study systems are increasingly being used to diagnose the sleep apnoea/hypopnoea syndrome, but validation is essential and detection of arousal's desirable. One such system (AutoSet) was validated on an event-by-event basis, and the hypothesis that sudden large breaths detected by this system mark arousal from sleep was also examined. Twenty consecutive patients (apnoea/hypopnoea index (AHI) 39+/-6 (SEM)) underwent polysomnography (PSG), which included real-time signals of AutoSet (Version 3.03) scored events. PSG respiratory events were defined using airflow and thoracoabdominal movement and AutoSet events using nasal pressure. All apnoeas were scored by both systems, but 41% more hypopnoeas were scored on PSG and these were clinically significant, with 78% ending in cortical arousal. Twenty per cent of apnoeas and hypopnoeas scored by the AutoSet occurred during wakefulness. Large breaths, defined as a two-thirds increase in ventilation, marked 77% of respiratory-associated but only 9% of spontaneous arousals. Large breaths also marked 48% of "autonomic" arousals following respiratory events without visible electroencephalographic changes. Twenty-seven per cent of large breaths occurred during wakefulness. This study shows that the AutoSet and the standard polysomnographic approach differ in their detection of hypopnoeas. The AutoSet underdetected hypopnoeas during sleep, but also included some hypopnoeas occurring during wakefulness. Detection of large breaths may potentially be useful for identifying respiratory arousals. Detection of periods of wakefulness may improve the accuracy of the system