Characterizing deep inspiration dynamics across obstructive sleep apnea severity levels

Obstructive sleep apnea (OSA) is characterized by recurrent upper airway obstructions during sleep. While the apnea-hypopnea index (AHI) is the standard for assessing OSA severity, it fails to capture physiological details in respiratory signals, particularly those related to airway responses to obstruction. This study investigates deep inspiration patterns following apneas and hypopneas across OSA severity levels to uncover airflow dynamics beyond what AHI reflects.

Using data from 202 participants in the Sleep Heart Health Study (SHHS), respiratory airflow signals were extracted and processed to isolate deep inspirations following respiratory events. Two standardization methods, interpolation and averaged random subsampling, were used to normalize time series lengths. Various interpolation techniques and orders for polynomial regression were evaluated to model individual and group-level respiratory patterns.

Post-event deep inspiration amplitude decreased progressively with increasing OSA severity, with severe OSA showing marked attenuation, particularly in the first and last sleep quartiles. Distance-based analyses confirmed prominent magnitude differences, while cosine and correlation metrics demonstrated preserved temporal shape, indicating impaired compensatory ventilation rather than altered recovery dynamics.

These results highlight that airflow signal morphology reveals respiratory compensation mechanisms in OSA. Post-event inspiratory patterns may serve as complementary markers of disease severity, offering a more detailed understanding of OSA beyond AHI.