A new oximetry algorithm for screening of sleep-disordered breathing

Pulse oximetry has been developed as a tool for non-invasive evaluation of arterial oxygen saturation. Recently, this method is also applied for overnight monitoring of sleep-disordered breathing (SDB). However, pulse oximetry has been not fully optimized for such a purpose. The algorithm to obtain the value of oxygen saturation included procedures to eliminate noises due to body movement, which could compromise the response of oximetry during a rapid change of oxygen saturation. In this study we aimed to evaluate the diagnostic ability of oximetry implemented with a new algorithm optimized for SDB as compared with that of oximetry with an old algorithm. Subjects are 57 patients who were referred for suspected sleep apnea. Ordinary polysomnography (PSG) with experimental oximetry recording was performed. The oximetry recorded all raw data in the memory which was analyzed later using the old and the new algorithm separately. The old algorithm has a filter characteristic corresponding to 5-second moving average and resolution of 1%, and new algorithm has a filter characteristic corresponding to 3-second moving average and resolution of 0.1%. The old algorithm saved the analyzed data every 5 seconds, while the new algorithm saved those every 1 second. We calculated the apnea-hypopnea index (AHI: number of apnea/hypopnea event per hour of sleep) from the PSG record and oxygen desaturation index (ODI: number of SpO₂-dip by 3% or more per hour of examination). The intraclass correlation coefficient between AHI and ODI was 0.70 for the old algorithm and 0.84 for the new algorithm. The diagnostic sensitivity/specificity to detect SDB patient with AHI≧15 when the ODI cutoff value of 10 was used were 0.81/0.95 and 0.97/0.76, for the old and the new algorithms, respectively. We conclude that the new algorithm provides oximetry with excellent response to apnea/ hypopnea events and oximetry implemented with this new algorithm is appropriate for the screening of SDB patients.