0481 Use of Continuous Wavelet Transform to Identify Central Sleep Apnoea from Overnight Oximetry Data

Abstract Introduction Overnight oximetry may be a useful tool to screen for sleep disordered breathing but, although reasonably specific, is not very sensitive. It cannot reliably distinguish central from obstructive sleep apnoea. The continuous wavelet transform is an advanced mathematical analysis technique that yields a high resolution time-frequency spectrogram that is able to give more information regarding the finer details of the oximetry. The aim of this study is to determine the differences in the continuous wavelet transform, between central and obstructive sleep apnea. Methods A retrospective analysis of the continuous wavelet transform of oximetry data extracted from a cohort of 209 diagnostic sleep studies was performed. The spectral characteristics of central sleep apnoea were compared to those of obstructive sleep apnoea and normal sleep results. Results The oximetry spectrogram for central sleep apnoea shows a clear dominant frequency of desaturations across all sleep stages, with a decrease in amplitude during REM sleep. In contrast the oximetry spectrogram in obstructive sleep apnoea shows no clear dominant frequency across the sleep stages, and an increase in the amplitude and frequency variability during REM sleep. There is a significant difference in the prominence of the maximum amplitude and in the variability of the frequency of the maximum amplitude in central sleep apnoea compared to obstructive (5.3 (3.1 - 15) vs 3.1 (2.5 - 5.1), p=0.001 and 1 (0.5 - 2.3) vs 1.5 (1.1–2.2), p=0.002 respectively). Conclusion The continuous wavelet transform is a mathematical technique that allows further information to be extracted from overnight oximetry data. There are clear differences in the oximetry spectrograms of central and obstructive sleep apnoea that can be used to distinguish between them. This can be further adapted to improve the utility of overnight oximetry as a screening tool. Support (If Any) No conflicts of interest or funding sources to declare.