Physiologic Phenotypes of Sleep Apnea Pathogenesis

Obstructive sleep apnea is a disorder characterized by recurrent episodes of pharyngeal obstruction during sleep (1), leading to repeated oxyhemoglobin desaturations and arousals from sleep. These acute physiologic perturbations can be characterized by polysomnographic metrics that quantify the degree of oxyhemoglobin desaturation and sleep fragmentation including the oxygen desaturation index, time spent with an Sa_O2 below 90%, respiratory disturbance indices (apnea and hypopnea indices) in non–rapid eye movement (non-REM) and REM sleep, and alterations in sleep architecture (arousal frequencies, sleep stage distribution, and transition indices). Increasing degrees of sleep fragmentation and oxyhemoglobin desaturation likely increase the patient's risk of developing clinical sequelae including neurocognitive dysfunction, metabolic dysregulation, and cardiovascular morbidity and mortality (2–5). Nevertheless, these metrics of sleep apnea severity provide little information about the underlying pathogenic mechanisms of this disorder.

Obstructive sleep apnea is due to increases in pharyngeal collapsibility during sleep (6). These increases can result from upper airway anatomic defects (7) and/or disturbances in neuromuscular control (6, 8). Structural defects have been linked to boney and soft tissue abnormalities, which load the pharynx and predispose to airflow obstruction during sleep. These defects are best summarized by measurements of pharyngeal critical closing pressures under conditions of absent or markedly reduced neuromotor tone (passive Pcrit) (9). Elevations in passive Pcrit have been associated with sleep apnea risk factors including obesity, central adiposity, male sex, and age (10, 11), suggesting its role as a physiologic mediator of sleep apnea susceptibility. As passive Pcrit increases, dynamic collapse of the pharynx limits inspiratory airflow to a maximal level (Vimax) as airway pressure downstream to the pharynx falls below Pcrit. Mechanical loads on the pharynx predispose to inspiratory airflow limitation, during which Vimax characterizes the severity of airflow obstruction during sleep (Figure 1) (12).

Figure 1.

Schematic illustrating intermediate physiologic phenotypes for sleep apnea pathogenesis. Pharyngeal mechanical loads elevate passive Pcrit (left panel). Dynamic collapse of the airway leads to the development of inspiratory airflow limitation (middle panel), as characterized by a plateauing of the inspiratory airflow waveform (I = inspiration, E = expiration) at a maximal level (Vimax) as effort increases (not shown). Compensatory neural responses (right panel) mitigate the fall in ventilation by increasing Vimax (upper right panel) or inspiratory duty cycle (lower right panel). Compensatory increases in mean inspiratory airflow (Vt/Ti) and inspiratory duty cycle (Ti/Ttot) restore minute ventilation, as described in the equation, where Vt is tidal volume, Ti is inspiratory time, and Ttot is the respiratory cycle period. E, expiration; I, inspiration.

Inspiratory airflow limitation can elicit neural responses that preserve ventilation during sleep (Figure 1) (8, 13, 14). Activation of pharyngeal muscles can decrease airway collapsibility (active Pcrit), which decreases the severity of airflow obstruction (see increased Vimax, Figure 1, top right) (6, 8, 15). Alternatively, individuals can compensate for inspiratory flow limitation by increasing the inspiratory duty cycle, as defined by the inspiratory time (Ti) over total respiratory period (Ttot) (see increased Ti/Ttot, Figure 1, lower right) (8, 16). At any given level of Vimax, lengthening the inspiratory duty cycle is the only means available to preserve ventilation when the pharynx flow-limits during sleep (17). As the inspiratory duty cycle increases, minute ventilation will increase accordingly, irrespective of respiratory rate and tidal volume (Figure 1, lower right) (17). Thus, neural responses can either mitigate or compensate for pharyngeal obstruction (flow limitation), helping to maintain ventilation during sleep.

Severe airflow obstruction can overwhelm these neural responses. What happens when these responses fail to protect the individual from developing obstructive sleep apnea? In the current issue of the Journal, Jordan and coworkers (pp. 1183) address this question by inducing airflow obstruction and monitoring the response (18). Airflow obstruction triggered an overshoot in ventilation with and without frank arousals from sleep. These investigators examined whether arousals further destabilized ventilation during sleep. Compared with trials that did not precipitate arousals, those with arousals were associated with greater degrees of ventilatory overshoot and a greater frequency of obstructive hypopneas. Of note, ventilatory instability ensued after arousals, regardless of the severity of airflow obstruction or associated decreases in pharyngeal muscle activity. These findings suggest that when neuromuscular responses fail to compensate for upper airway obstruction during sleep, arousals contribute to the development of respiratory instability and help to propagate recurrent obstructive apneas and hypopneas.

Jordan's study suggests that ventilatory responses to arousals constitute an independent physiologic phenotype of sleep apnea susceptibility. If so, we should extend our ability to characterize physiologic phenotypes in every sleep recording. If suitable technology were adopted to quantify airflow accurately, ventilatory metrics including Vimax, timing indices, and minute ventilation could be determined during sleep, respiratory-triggered arousals, and sleep-wake transitions. Moreover, robust physiologic markers of sleep-wake stability could be used to quantify effects of arousals on ventilatory stability. Current evidence suggests that distinct neural mechanisms govern sleep-wake stability (19). Sleep-disordered breathing episodes and sleep-wake propensities can interact, triggering alterations in sleep architecture, arousals, sleep-wake transitions, and α intrusion during sleep (20). Much work remains to discern which metrics predict ventilatory and sleep-wake instability once compensatory neural responses fail to stabilize ventilation during sleep.