Rationale: In the “classic” model of obstructive sleep apnea (OSA), apneas and hypopneas are characterized by a loss of airway patency at sleep onset that precipitates a reflex rise in neural ventilatory drive and ultimately arousal from sleep. In direct contrast, a loss of neural ventilatory drive—and attendant loss of dilator muscle tone—appears to be a causal determinant of respiratory events for many patients (“drive-dependent” model). Here we determine the extent to which the classic vs drive-dependent models of OSA are best supported by physiological measurement.

Methods: In N=50 OSA patients (5 to 91 events/hr), we recorded ventilation (oronasal mask and pneumotach) and ventilatory drive (per calibrated intra-esophageal diaphragm EMG) during overnight polysomnography. Data were analyzed at an individual event level and through ensemble-averaging respiratory events (signals synchronized at respiratory event termination). Patients were scored visually as “drive-dependent” if ventilation fell/rose during events simultaneously with ventilatory drive (not necessarily equal magnitude). Measures of the relative loss of ventilation and ventilatory drive and the correlation between these signals captured the extent to which a fall in drive potentially explains the ventilatory decline during observed events.

Results: On average, ventilation and ventilatory drive fell during events by 68[28,100]seline and 20[3,42]seline respectively (median [IQR]). The time course of ventilation and ventilatory drive during events was strongly and positively correlated (R = 0.78[0.24,0.94]), with nadir ventilation occurring just 5[2,8]s after nadir ventilatory drive. Visual scoring indicated that 30/50 (60%) of participants exhibited a “drive-dependent” phenotype; compared to the “classic” phenotype, drive-dependent patients exhibited dips in drive that were more closely correlated with ventilation (0.92[0.85-0.95] vs 0.13[−0.38-0.41]) and more closely aligned (2[0-3] vs 8[6-14] s), supporting the visual scoring.

Conclusions: In contrast to the “classic” pattern of OSA, ventilatory drive typically falls during obstructive respiratory events suggesting that airflow loss may be causally dependent on lowered ventilatory drive in most patients. Novel treatments that prevent dips in drive may be effective in a greater proportion of patients than previously appreciated.