Kidney Week

Abstract: PO0939

Monitoring of Intradialytic Sleep Apnea in Hemodialysis Patients

Session Information

• Leveraging Technology and Innovation to Predict Events and Improve Dialysis Delivery
  November 04, 2021 | Location: On-Demand, Virtual Only
  Abstract Time: 10:00 AM - 12:00 PM

Category: Dialysis

• 701 Dialysis: Hemodialysis and Frequent Dialysis

Authors

• Paneque Galuzio, Paulo, Renal Research Institute, New York, New York, United States

Paulo Paneque Galuzio, PhD

• Cherif, Alhaji, Renal Research Institute, New York, New York, United States

Alhaji Cherif,

• Tao, Xia, Renal Research Institute, New York, New York, United States

Xia Tao,

• Thwin, Ohnmar, Renal Research Institute, New York, New York, United States

Ohnmar Thwin,

• Thijssen, Stephan, Renal Research Institute, New York, New York, United States

Stephan Thijssen,

• Kotanko, Peter, Renal Research Institute, New York, New York, United States

Peter Kotanko,

Background

Breathing disorders are frequent in end-stage kidney diseases, with more than 50% of hemodialysis (HD) patients experiencing sleep apnea syndrome (SAS). SAS is associated with lower health-related quality of life and represents a significant cardiovascular risk factor. HD patients with SAS are at greater risk of mortality. The study aimed to monitor intradialytic SAS in HD patients using oxygen saturation (SaO₂) measurements.

Methods

Crit-Line® monitor was used to record SaO₂ at 1 Hz, for 2 HD sessions with a mean duration of 3.5±0.5 h. For each patient, we calculated: oxygen desaturation density (ODD₁₀), which counts 3% drops in SaO₂ from a baseline for at least 10 s long; a10^th order permutation entropy to quantitate complexity; and optimal recurrence threshold (ε_optimal) to account for dynamic variations and degree of predictability in the SaO₂. These quantities were subjected to machine learning methods to predict intradialytic SAS, as quantified by the ODD₁₀ value and the SAS classification by the American Sleep Disorders Association.

Results

We examine intradialytic SAS severity in 16 patients (age of 54±11 years, 63% males, 69% Black) with arteriovenous vascular access. Mean SaO₂ was 94.3±2.1%. Figure 1A shows a typical SaO₂ annotated with the SAS intensity assessed by ODD₁₀ (Fig. 1B). The two calculated metrics are plotted in Figs. 1C-D. The results reveal dynamic characteristic patterns of SAS with differential severity scores during HD. Figures 1E-G show the ROC for the classifiers when considering episodes of at least mild, moderate, or severe SAS, respectively. The maximum AUC is 0.93 for severe SAS episodes.

Conclusion

Our analysis suggests that entropy and recurrent-based quantifiers could be used as predictive indicators of intradialytic SAS. However, further studies are needed to assess their relationships to clinical outcomes.