The use of combined thermal/pressure polyvinylidene fluoride film airflow sensor in polysomnography

Abstract

The technologies recommended by the American Academy of Sleep Medicine (AASM) to monitor airflow in polysomnography (PSG) include the simultaneous monitoring of two physical variables: air temperature (for thermal airflow) and air pressure (for nasal pressure). To comply with airflow monitoring standards in the sleep lab setting thus often requires the patient to wear two sensors under the nose during testing. We hypothesized that a single combined thermal/pressure sensor using polyvinylidene fluoride (PVDF) film responsive to both airflow temperature and pressure would be effective in documenting abnormal breathing events during sleep. Sixty patients undergoing routine PSG testing to rule out obstructive sleep apnea at two different sleep laboratories were asked to wear a third PVDF airflow sensor in addition to the traditional thermal sensor and pressure sensor. Apnea and hypopnea events were scored by the sleep lab technologists using the AASM guidelines (CMS option) using the thermal sensor for apnea and the pressure sensor for hypopnea (scorer 1). The digital PSG data were also forwarded to an outside registered polysomnographic technologist for scoring of respiratory events detected in the PVDF airflow channels (scorer 2). The Pearson correlation coefficient, r, between apnea and hypopnea indices obtained using the AASM sensors and the combined PVDF sensor was almost unity for the four calculated indices: apnea-hypopnea index (0.990), obstructive apnea index (0.992), hypopnea index (0.958), and central apnea index (1.0). The slope of the four relationships was virtually unity and the coefficient of determination (r (2)) was also close to 1. The results of intraclass correlation coefficients (>0.95) and Bland-Altman plots also provide excellent agreement between the combined PVDF sensor and the AASM sensors. The indices used to calculate apnea severity obtained with the combined PVDF thermal and pressure sensor were equivalent to those obtained using AASM-recommended sensors.