0682 Recognition of High Loop Gain Respiration: Comparison of Self-Similarity Analysis and Clinician Interpretation

Introduction: High Loop Gain (HLG) respiration noted on diagnostic Polysomnography (PSG) or during positive airway pressure titration is a marker of treatment complexity for Obstructive Sleep Apnea (OSA). Self-Similarity (SS) software analysis of breathing cycles occurring in association with respiratory events has been shown to be an accurate and clinically useful tool to detect HLG; however, little data exist regarding identification by sleep medicine physicians. Our goal was to compare SS scores with clinician interpretation of periodic breathing or central sleep apnea.

Methods: SS analysis was run on a selection of polysomnograms from a data archive with a calculated central Apnea Hypopnea Index (SScAHI) in addition to %SSany index representing self similarity in breathing cycles. Demographic data and basic PSG values including AHI, and oxygenation data were recorded from the archived reports. Clinician interpretations from the archived studies were also examined for identification of key terms including “periodic breathing” and “high loop gain”. Chi square analysis was performed to examine the relationship between clinician recognition of HLG and %SSany.

Results: Data from 148 patients (81 female, 67male) (mean ± SD) age 58.2 ± 18.2 years with BMI 29.7 ± 7.4. PSG revealed mean AHI (18.0 ± 15.4), cAHI (2.6 ± 6.0). SS analysis revealed mean %SS 10.9 ± 11.2. When stratified across 3 separate categories of SS 5% (Pearson chi2 = 9.50 Pr = 0.002), 10% (Pearson chi2 = 5.59 Pr = 0.018) and 20% (Pearson chi2 = 16.1509, Pr = 0.000) clinician interpretation lacked mention of HLG. Taking into account cAHI at even the SS 20% category also failed to elicit recognition of high loop gain (Pearson chi2 = 5.07, Pr = 0.024).

Conclusion: High Loop Gain respiration is often overlooked in clinician interpretation of polysomnography, and thus the potential benefit of adjunctive therapies including pharmacotherapy and CO2 modulation which target HLG are being lost. This highlights the benefits of methods such as SS analysis which can be used to detect HLG independent of clinician recognition of breathing cycle patterns.