Pulse wave amplitude reduction: a surrogate marker of micro-arousals associated with respiratory events occurring under non-invasive ventilation?

INTRODUCTION:Respiratory events occurring under non-invasive ventilation (NIV) may produce sleep fragmentation. Alternatives to polysomnography (PSG) should be validated for providing simple monitoring tools for patients treated at home with NIV. OBJECTIVES:To study the value of pulse wave amplitude (PWA) reduction as a surrogate marker of cortical micro-arousals associated with respiratory events occurring during NIV. METHODS:27 PSG tracings under NIV recorded in 9 stable patients with Obesity Hypoventilation Syndrome (OHS), under 3 different ventilator modes (no back-up rate, low or high back-up rate) were analyzed. For all respiratory events (obstructive, central, or mixed event), the association with EEG-micro-arousals, PWA reduction of more than 30% and the presence of associated SpO2 desaturation ≥ 4% was recorded. RESULTS:2474 respiratory events during NREM sleep were analyzed. 73.6% were associated with an EEG-MA, 91.4% with a ≥ 4% decrease in SpO2, and 74.9% with a significant PWA reduction. Sensitivity of PWA for the detection of an EEG-micro-arousal related to a respiratory event was 89.1% [95%CI: 76.7-95.3]. Positive predictive value (PPV) was 87.0% [95%CI: 75.0-94.0]. Sensitivity of PWA was highest in the S mode, compared to both other S/T modes, p = <0.001. Sensitivity of PWA was also higher for central and mixed events, compared to obstructive respiratory events, p = <0.05. CONCLUSIONS:PWA reduction is a sensitive marker with a high PPV for the detection of EEG-MA associated with respiratory events during NREM sleep in stable OHS patients treated by NIV. In this situation, PWA could be used to improve scoring of hypopneas and allow an appropriate assessment of sleep fragmentation related to respiratory events.