Corralation between desaturation indices of oxygen saturation variability in severe obstructive sleep apnea: a pilot study

Abstract

Background: Obstructive sleep apnea (OSA) is a respiratory disorder characterized by recurrent airflow obstruction caused by total or partial collapse of the upper airway. The aim of this study is to compare the accuracy of the different indices derived from pulse oximetry using free software in the diagnosis of severe OSA and to determine the relationship of these indices with the apnea-hypopnea index (AHI), cumulative time spent below 90% SpO₂ (T90) and the oxygen desaturation event index (ODI).

Methods: Cross-sectional study with 20 untreated severe OSA patients and 11 subjects with AHI<5 referred to a sleep laboratory. The AHI, T90 and ODI indices were calculated from pulse oximetry simultaneously during polysomnography. To analyze variability of SpO₂ in each pulse wave  (SpO₂V), interval of oxygen pulse wave (IOP) index, standard deviation of IOP (STDIOP), mean of oxygen saturation (MOS), root mean square of IOP standard deviation (SDIOP) of SpO₂V were calculated.

Results: OSA subjects presented with higher values of IOP, STDIOP, SDIOP when compared to controls (p<0.05). The area under the ROC curve of IOP, STDIOP, SDIOP were 0.93, 0.99, and 0.99, respectively. T90 demonstrated a positive correlation with IOP (r=0.94), STDIOP (r=0.64) and SDIOP (r=0.73) as well as ODI with IOP (r=0.79), STDIOP(r=0.78), SDIOP (r=0.73) and AHI with STDIOP (r 0=76), SDIOP (r=0.73).

Conclusions: SpO₂V was correlated with AHI, ODI and T90 indices and presented positive predictive value for OSA. The SpO₂V could be useful and allows for a non-invasive widely applicable diagnostic technique for severe OSA.