MO779DETECTION OF INTRADIALYTIC SLEEP APNEA IN HEMODIALYSIS PATIENTS

Abstract

Listen

Translate article

Abstract Background and Aims Sleep apnea syndrome (SAS) has emerged as an important cardiovascular risk factor in hemodialysis (HD) patients. In patients with SAS, high-frequency oscillations of arterial oxygen saturation (SaO2) and arousal can be used to characterize repetitive respiratory cessation. Although these repetitive patterns have been observed in SAS patients, it has not been studied during HD treatment. In this study, we aim to detect intradialytic SAS in HD patients to provide novel prognosticators of adverse events related to SAS. Method We use the Crit-Line® monitor to record SaO2 at 1 Hz frequency, and video recording to capture periods of wakefulness for the entire treatment session (3.5±0.5hr). We developed a recurrent-based time-series analysis to characterize the desaturation and arousal in SaO2. For each patient, we compute four metrics, namely: (i) the probability of recurrence for any given state (recurrent rate, RR), (ii) systemic predictability (determinism, DET), (iii) occurrence of laminar state (LAM), and (iv) permutation entropy to quantify complexity. These four quantities were used to detect episodes of SAS in patients while undergoing HD treatments. Results We study 16 HD patients with arterio-venous vascular access. Their age was 54±11 years, 63% were males and 69% were African Americans. SaO2 was measured during two HD sessions per patient. The mean SaO2 was 94.3±2.1%. We observe distinguishable patterns in which periods of awake and asleep are differentiated. Figure 1 shows a typical SaO2 annotated with the period of awake and asleep (shaded area in Fig. 1A), during which the four metrics decrease during the corresponding period where the patient is asleep (Fig. 1B-C). Figure 1D-E shows the consistency of the metrics in all 16 patients, in which DET and entropy are analyzed around the onset of sleep. We observe monotonically decreasing trends around the episode of sleep, suggesting that these results are sensitive to the sleep states, and potentially to an episode of intradialytic SAS. Conclusion Our preliminary analysis suggests that DET, LAM, RR, and entropy may contain characteristic properties that could be used to detect the onset of intradialytic SAS, which are characterized by the high-frequency oscillatory patterns in SaO2. Further studies are needed to identify SAS-related desaturation and arousal detected in the SaO2 measurement. The predictive power of these features and their relationships to clinical outcomes warrants further investigation.