A reminder that experimentally induced intermittent hypoxia is an incomplete model of obstructive sleep apnea and its outcome measures

For two to three decades many scientists and clinicians in the field of sleep medicine believed that exposure to intermittent hypoxia leads to numerous detrimental outcomes, including cardiovascular, metabolic, and cognitive dysfunction (3). The link between intermittent hypoxia and detrimental outcomes was thought to exist independent of the perturbation (i.e., sleep apnea, experimentally induced hypoxia) and the pattern, intensity, and duration of exposure. Within the past decade a detailed understanding and broader acceptance of the physiological effects of intermittent hypoxia has emerged (12, 13, 17, 21). It is now recognized that protective or detrimental outcomes following exposure to intermittent hypoxia are possible and in many cases dose dependent (12, 13, 17, 21). However, there remains in the field some lack of understanding or acknowledgment that experimentally induced intermittent hypoxia is not a representative model of sleep apnea, which is characterized by a number of hallmarks, including intermittent hypoxia.

Intermittent hypoxia, administered as an experimental intervention, is initiated by reducing the fractional concentration of inspired oxygen or reducing atmospheric pressure. Administration of the stimulus to humans or animals typically occurs in the absence of other perturbations that accompany sleep apnea (15). These perturbations include, but are not limited to, large swings in intrathoracic pressure against a closed airway, persistent arousal from sleep following the termination of a breathing event, and carbon dioxide measures that oscillate between hypocapnia and hypercapnia (3, 18). In addition, the application and characteristics of experimentally induced intermittent hypoxia often differ from the characteristic pattern of exposure to intermittent hypoxia that accompanies sleep apnea (15). In animal and human models, experimentally induced hypoxia is frequently introduced in a square wave manner, with changes in the fractional concentration of oxygen occurring abruptly at the onset and offset of the episode (10, 15). Likewise, the typical duration of exposure (e.g., 2–5 min) is longer than episodes that occur as a consequence of apneic events (e.g., 20 s) (15). Thus the hypoxic profile may not simulate that observed during obstructive events in humans. However, it should be noted that a few studies have successfully modeled the hypoxic profile in rodents using experimentally induced hypoxia (9, 10). In addition, chronic exposure to intermittent hypoxia in animal models is often administered without controlling for arousal state (i.e., wake versus sleeping) (15). This is not the case in those with sleep apnea, since exposure to intermittent hypoxia only occurs during sleep. Consequently, the characteristics of the model frequently employed to examine the impact of experimentally induced intermittent hypoxia are often difficult to compare with a human model of sleep apnea. Instead, the utility of animal and human models exposed to experimentally induced hypoxia likely reside in determining outcomes associated with various doses of intermittent hypoxia. In other words, varying the duration, intensity, and pattern of exposure will eventually lead to establishing the appropriate dose and approach to induce protective physiological outcomes. This could lead to further promotion of intermittent hypoxia as a therapeutic modality (12, 13, 17, 21). On the other hand, presuming that a physiological response to a typical experimentally administered dose of intermittent hypoxia has general applicability to humans with obstructive sleep apnea is less certain.

Thus, examining the effects of intermittent hypoxia, induced by obstructive sleep apnea, on various comorbidities requires an experimental model that incorporates airway obstruction, accompanied by variations in intrathoracic pressure swings, arousal from events, and oscillations in carbon dioxide measures between hypocapnia and hypercapnia (3, 18). As a starting point animal models likely provide the best opportunity to examine these complex interactions in a controlled manner (2). There are numerous models available that might be employed to explore if mild versus moderate versus severe intermittent hypoxia in the presence of apneic events leads to protective versus detrimental responses. In the past, dog (7, 8), lamb (4), and pig (11) models, which nicely modeled the human condition, were employed. Continued use of these large animal models, at least in some laboratories, may provide important mechanistic insight because of their effectiveness in simulating sleep apnea. However, recent use of large animals to model obstructive sleep apnea has been infrequent, principally for economic reasons. Instead, rats and mice are now used as experimental models (1), and these models can be divided into two categories.

One category comprises rat or mouse models that display spontaneous apneas. Apneas recorded from these animals are typically central in nature, and consequently obstructive sleep apnea is not modeled (1, 2). Despite this limitation, the model might provide important mechanistic insight into the relationships between the absence of respiratory motor activity (i.e., central apnea), the concurrent intermittent hypoxia, and various forms of respiratory plasticity initiated by these stimuli. Indeed, recent findings in rats have shown that central apneic events may initiate a form of respiratory plasticity referred to as inactivity-induced inspiratory motor facilitation (5, 14). If this model is employed to explore the impact of central apneic events, the role that pattern, intensity, and duration of intermittent hypoxia has on outcome variables may be difficult to decipher, since events are random and variable.

There are also models in which the upper airway can be obstructed in a controlled fashion (1, 2). Consequently, the number and characteristics (i.e., intensity, duration, pattern) of events are controlled variables. These models, combined with measures of electroencephalography to monitor the presence and intensity of arousal, would provide an opportunity to explore how intermittent hypoxia interacts with other hallmarks (i.e., arousal) of sleep apnea to induce a characteristic outcome. The model could be used, for example, to determine if mild intermittent hypoxia elicits protective outcomes in the presence of significant arousal from sleep. A model capable of controlling apneic events might also be combined with controlled models of sleep fragmentation (6, 16, 19, 20). Thus, both obstruction of the airway and arousal could potentially be controlled. As a result, a powerful tool to explore the impact of intermittent hypoxia initiated by sleep apnea could emerge. The findings could fill a present void in our understanding as to how the complex interaction of various hallmarks of sleep apnea initiates outcomes that might be protective or detrimental depending on the interaction among the variables.

GRANTS

This work was supported by awards from the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development (I01CX000125 and 15SRCS003), and from the National Heart, Lung and Blood Institutes (R56-HL-142757 and R01-HL-142757).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author.

AUTHOR CONTRIBUTIONS

J.H.M. drafted, edited, and approved the final version of the manuscript.