Chest

Chest

Volume 123, Issue 3, March 2003, Pages 704-710
Journal home page for Chest

Clinical Investigations
Sleep and Breathing
Evaluation of Unattended Automated Titration To Determine Therapeutic Continuous Positive Airway Pressure in Patients With Obstructive Sleep Apneaa

https://doi.org/10.1378/chest.123.3.704Get rights and content

Background

Determination of the therapeutic pressure during continuous positive airway pressure (CPAP) therapy is usually performed by a technician during polysomnography. In recent years, several devices for automated adjustment of the therapeutic pressure by the means of computerized algorithms were developed. The aims of the present study were to compare two different devices for automated titration and to verify if unattended automated titration is a feasible strategy to determine the therapeutic CPAP.

Methods

We enrolled 16 consecutive patients with obstructive sleep apnea syndrome (OSAS) defined by an apnea-hypopnea index > 20/h. Automated titration was performed in the hospital using two CPAP devices (Autoset; Resmed; North Ryde, Australia; and Somnosmart; Weinmann; Hamburg, Germany) in random order for 2 consecutive nights, based on different signals for the detection of respiratory events. During titration, there was no direct supervision by a technician, and polysomnography was not recorded. We defined the therapeutic pressure as the 95th percentile of the airway pressure over time (P95).

Results

We observed significant differences of the P95 between the two devices, with an average of 7.0 ± 2.5 cm H2O for the Somnosmart and 9.9 ± 2.6 cm H2O for the Autoset (p = 0.005) [mean ± SD]. There was a considerable lack of agreement between the two devices, with a bias of 3.0 cm H2O and limits of agreement ranging from + 9.3 to − 3.2 cm H2O. We found no significant correlation between the paired differences of P95 and either indexes of severity of OSAS or lung function variables.

Conclusion

Automated titration based on the analyses of flow (Autoset) or forced oscillations (Somnosmart) predicted significant different therapeutic pressures for fixed CPAP therapy. Thus, unattended automated titration performed during 1 night of hospital stay with commercially available devices cannot be used to determine accurately the therapeutic CPAP in patients with OSAS.

Section snippets

Patients

The sample population consisted of consecutive patients referred for suspicion of OSAS. However, since we usually perform sleep studies in the pneumology unit, the incidence of respiratory diseases (associated COPD, obesity-hypoventilation syndrome) in our patients is quite high. All patients underwent complete evaluation, including a full polysomnography, arterial blood gas analysis, and lung function testing.

Patients were included in the present study if they met the following criteria: (1)

Results

We included 16 consecutive patients (13 men and 3 women) who underwent standard polysomnography in our department. Demographics and lung function parameters are reported in Table 1. Mean age was 60 ± 10 years. Most of the patients were obese, since the median body mass index was 36 (interquartile range [IQR], 32 to 41).

Lung function variables demonstrated frequent association with COPD with various degrees of airways obstruction. Mean FEV1 was 71 ± 22% of predicted for the group. The median FEV1

Discussion

The present study demonstrates clearly that for a given patient, the recommended pressures (P95) obtained by two different autotitrating CPAP devices can differ considerably. Indeed, the difference of P95 was as high as 8 cm H2O in one patient. The P95 pressures and the P50s were significantly higher for the Autoset device compared to the Somnosmart device (p = 0.005 for each comparison). There was a great bias of 3 cm H2O and a complete lack of agreement between the two devices (limits of

Conclusion

Automated titration based on the analysis whether of flow (Autoset) or of forced oscillations (Somnosmart) predicted significant different therapeutic pressures for fixed CPAP therapy in patients with OSAS. These differences in recommended pressures between the Autoset and the Somnosmart were not due to differences in the severity of OSAS nor to the presence of an underlying lung disease. They could be explained by differences in algorithms, different analyses of signals, and a possible effect

References (13)

There are more references available in the full text version of this article.

Cited by (0)

View full text
Copyright © 2003 The American College of Chest Physicians. Published by Elsevier Inc. All rights reserved.