Introduction: Automatic continuous positive airway pressure (CPAP) devices are increasingly being
used for treating patients with obstructive sleep apnoea syndrome (OSAS). Most of
these devices are based on detecting respiratory events from the analysis of the breathing
flow shape and/or snoring. However, other devices also include the detection of upper
airway resistance increases for identifying obstructive events. The differences in
the detection method of these devices may influence their strategy for modifying the
applied CPAP during treatment, specifically in the presence of central and obstructive
events. The aim of this bench study was to evaluate the response of different commercial
automatic CPAP devices to simulated respiratory events with and without increases
in upper airway resistance during sleep. Methods: A patient simulator was used for evaluating the automatic CPAP devices in the bench
in a systematic and controlled way. The patient simulator was able to reproduce any
type of flow waveform (Farré et al. Am J Respir Crit Care Med 2002; 166: 469–73) and
included a servocontrolled valve capable of reproducing precisely the increases in
airway resistance typically found in patients with OSAS (J. Rigau et al. Eur Respir
J 2003; 22(45): 181s). Three devices with detection algorithms based on the flow shape
and snoring were analyzed: REMstar auto (Respironics, USA), Autoset Spirit (Resmed,
Australia), PV10i (Breas, Sweden). Two additional devices, which detect also increases
in airway resistance, were analyzed: SOMNOsmart 2 (Weinmann, Germany) and Autoset
II Plus (ResMed, Australia). Each device was connected to the patient simulator and
was submitted to hypopnoea flow patterns with different flattened inspiratory contours
and apnoea flow patterns with simultaneous increases in the airway resistance (obstructive
hypopnoeas and apnoeas) and without increases in airway resistance (central hypopnoeas
and apnoeas), continuously repeated regardless of the device reaction. Results: In the presence of obstructive apnoeas with a resistance of around 75 cmH2O s/L,
all the devices modified the applied pressure at different rates. The time required
for each device for increasing the pressure from 4 to 10 cmH2O was 4min for Autoset
II Plus, 5min for Autoset Spirit, 7min for PV10i, 12min for REMstar Auto and 13min
for SOMNOsmart 2. When the same apnoea pattern was reproduced without an increase
of airway obstruction (central apnoeas), the Autoset Spirit, the PV10i and the REMstar
Auto devices increased the applied CPAP exactly in the same way as during the obstructive
apnoeas. However, the SOMNOsmart 2 and the Autoset II Plus, did not modify the pressure,
keeping the initial CPAP of 4 cmH2O. The different flattened inspiratory flow shapes
were evaluated by the devices in a different way leading to pressure changes in a
pattern specific subgroup of devices only. The pressure reaction was independent of
simulated partial airway obstructions in the devices without resistance detection
and also in Autoset II Plus, since this device detects obstructions only during apnoeas.
The SOMNOsmart2, which detects airway resistance continuously, altered its pressure
response when no obstruction was present. Discussion: The automatic CPAP devices including the detection of increases in the upper airway
resistance are able to differentiate central from obstructive respiratory events.
This allows avoiding an unnecessary increase in the applied CPAP in the presence of
central events characterised by an absence of airway obstruction.
Supported in part by Measure, Check & Control, GmbH & Co. KG.
