Internal Medicine
Online ISSN : 1349-7235
Print ISSN : 0918-2918
ISSN-L : 0918-2918
Predicting Continuous Positive Airway Pressure from a Modified Split-Night Protocol in Moderate to Severe Obstructive Sleep Apnea-Hypopnea Syndrome
Ming-Lung ChuangI-Feng LinJanine R.E. VintchYu-Fang Liao
Author information

2008 Volume 47 Issue 18 Pages 1585-1592


Background Most split-night sleep (SNS) studies have enrolled Caucasian patients with moderate-to-severe obstructive sleep apnea-hypopnea syndrome (OSAHS), with different apnea-hypopnea index (AHI) criteria for diagnosis and successful CPAP titration from the standard full-night study (FNS). This study aimed to describe a shortened diagnostic SNS in the Taiwanese population using diagnostic and CPAP therapeutic criteria identical to the standard FNS, and to generate a predictive equation for effective CPAP (Peff) titration.
Methods A total of 418 patients with moderate-to-severe OSAHS were enrolled in this study. The duration of the diagnostic SNS was shortened to approximately 1-2 hours. The diagnostic accuracy and the success rate of CPAP titration in SNS were evaluated and a statistical model for Peff was built. The accuracy of the predictive equation was validated in another 127 patients.
Results Eighty-nine percent of patients with moderate-to-severe OSAHS diagnosed in the preceding FNS yielded the same grade of OSAHS in SNS. These reproducible patients were more severe, with 88% attaining successful CPAP titration in SNS. The predictive Peff=1.98+0.184×BMI+0.01×AHISNS +0.016×DISNS (R2=0.28, p<0.0001) where BMI was the body mass index and DI the desaturation index. The | predictive Peff- Peff | was within 2 cm H2O in 84% of the study group and 73% of the validation group.
Conclusion The modified split-night protocol and the predictive equation for CPAP can be useful in Taiwanese patients with moderate-to-severe OSAHS. Our findings may shorten the waiting time for polysomnography.

Information related to the author
© 2008 by The Japanese Society of Internal Medicine
Previous article Next article