1. Periodic breathing is known to be associated with cyclic fluctuations in heart rate. The purpose of this study was to evaluate the capability of spectral analysis of heart rate variability to identify episodes with periodic breathing in patients suspected of having sleep apnoea syndrome.
2. Forty-eight subjects complaining of chronic daytime sleepiness were studied using polysomnography and additional monitoring of Holter-ECG and synchronized pulse oximetry. The recordings were divided into 20 min episodes which were identified as recordings registered during normal breathing, periodic breathing, and periods of both normal and abnormal breathing. Power spectral analysis was performed on episodes which met the criteria of stationarity of data (313 episodes with normal breathing, 264 episodes with continuous periodic breathing, 80 episodes with both normal and periodic breathing pattens).
3. The ability of parameters, derived from analysis of heart rate variability, to discriminate between episodes with normal and periodic breathing was assessed by receiver-operating characteristic analysis.
4. The spectral power component in the frequency range 0.01–0.07 Hz revealed the greatest accuracy for discriminating between normal and periodic breathing (area under the receiver-operating characteristic curve = 0.929; standard error = 0.009). The analysis of the episodes classified as false-positive at a given test sensitivity of 90% and a corresponding specificity of 77% revealed that half of these episodes had been recorded during transient central nervous arousal reactions related to periodic leg movements or heavy snoring.
5. We concluded that power spectral analysis of heart rate variability offers a possible means of identifying episodes of sleep-related breathing disorders or periodic leg movements. Therefore, analysis of heart rate variability may be a valuable additional diagnostic tool in patients undergoing Holter-ECG recording.