Upper Airway Myopathy is Important in the Pathophysiology of Obstructive Sleep Apnea

In this article, the term “myopathy” will encompass any process which adversely impacts on upper airway muscle force generation and resultant tissue displacement. Key evidence which supports a role for upper airway myopathy in the pathophysiology of OSA will be highlighted.

OSA is associated with changes in contractile function of upper airway dilator muscles. In order to adapt to increased contractile demands, skeletal muscle fiber phenotype can undergo modification from a shift from oxidative slow-twitch, fatigue resistant Type I, to glycolytic fast-twitch Type II fibers, which generate increased force but are more prone to fatigue. Several groups have evaluated upper airway muscle histological, biochemical and in vitro contractile properties in OSA. Series et al ^1–3 studied musculus uvulae and found an increased proportion of fast-twitch type IIa fibers along with augmented levels of enzymes associated with anaerobic metabolism in OSA patients vs. snorers¹. An increase in type IIa fiber prevalence was also found in the genioglossus muscle of OSA patients². Contractile characteristics of musculus uvulae but not genioglossus were studied in vitro^1–3. Force generation was slightly higher for musculus uvulae in OSA than control^1,3 but muscle fatigability showed a nonsignificant tendency to be greater in OSA patients¹. Smirne et al⁴ also reported a type IIa fiber shift. Carrera et al^5,6 studied the characteristics of genioglossus in OSA patients vs controls and reported a type II fiber shift, and showed preserved tetanic force generation, but in nonobese OSA patients, clearly increased muscle fatigability.

Changes similar to those in humans have been reported in animal models. Petrof et al⁷ reported Type II fiber shifts in the sternohyoid and geniohyoid muscles in the English bulldog, a spontaneous model of OSA. In the rat, both short- and longer-term intermittent hypoxia^8,9 have been reported to produce type II fiber shifts, preserved contractility, but increased fatigability of geniohyoid and sternohyoid muscles.

Thus a theme emerges that upper airway muscle loading and/or hypoxia in OSA leads to shifts to fast-twitch but more fatigue-prone fiber types, and that in vitro muscle contractility is preserved or even slightly increased but at the cost of increased fatigability, i.e. a reduced ability to sustain force generation with repeated stimulation. While these changes likely represent an adaptive response to increased upper airway loading,¹⁰ there is a high likelihood that in vivo, the increased susceptibility to fatigue could lead to a progressive loss of force generation, muscle shortening and tissue displacement by upper airway dilators during the repeated high-level activation required to restore airway patency^7,10. Indeed, in vivo changes in muscle function with recurrent apneas could be a contributing factor to the progression of OSA severity across the night that we have previously reported^11,12.

There is evidence for dysfunctional mechanical coupling of upper airway muscles in OSA, which attenuates the tissue displacement resulting from dilator muscle contraction. In an important study, Series and colleagues³ measured the elastance of uvular tissue (which incorporated the musculus uvulae). They found that uvular elastance was increased in OSA patients, such that the degree of shortening of intact uvular tissue specimens for a given force of stimulated muscle contraction was significantly less than for nonapneic controls. They subsequently dissected out the musculus uvulae and demonstrated that muscle force generation per se was again slightly greater for OSA than control^1,3. Thus even if the force generating capacity of the upper airway dilator muscles is preserved, the effects of muscle activation on tissue displacement in OSA are significantly reduced, likely due to connective tissue alterations³. This would therefore be expected to attenuate the effects on upper airway patency of dilator muscle activation.

There is strong evidence for upper airway dilator muscle denervation in OSA patients, which would impair muscle contractile function in vivo. We have described an oropharyngeal sensory impairment in OSA^13–15 which has been confirmed by others^16,17 and which is partially reversible with CPAP¹³. We have also identified a similar impairment at the laryngeal level which correlates with OSA severity.^14,15 Woodson et al¹⁸ described nerve demyelination in upper airway tissue, and Friberg et al¹⁹ reported changes in upper airway mucosal nerve endings consistent with a pattern of injury and repair. These findings support the presence of an upper airway sensory neuropathy in OSA, which is proposed to be due to mechanical trauma associated with snoring and apneas (vibration, tissue traction, etc.) during obstructive events, oxidative stress related to hypoxia-reoxygenation, and inflammation related to both of these factors.^{13–15,20–22}

There is every reason to believe that this neuropathic process would not be isolated to sensory nerves, but would also affect motor nerves, and thereby lead to muscle denervation. Indeed, there is growing morphologic and physiologic evidence for upper airway muscle denervation in OSA. Immunohistochemical studies from several groups have demonstrated classic findings of muscle denervation including characteristic fiber type grouping, grouped atrophy, and increased fiber size variability (fiber atrophy and hypertrophy),^20,23–25 particularly in palatal muscle specimens obtained from OSA patients as compared with control. Our laboratory has provided further immunohistochemical evidence for ongoing, active denervation in OSA palatal muscle in that we identified increased expression of Neural Cell Adhesion Molecule, a subsarcolemmal protein which is transiently expressed in denervated muscle cells.²¹ There is also physiologic evidence for upper airway muscle denervation in OSA. Svanborg has described a high prevalence of characteristic denervation potentials in palatal muscles of OSA patients.²⁰ More recently, an increase has been observed in OSA patients vs. controls in the area and duration of single motor unit action potentials recorded in genioglossus during quiet breathing (Gandevia, Saboisky, and colleagues, unpublished observations). These findings are characteristic for denervation. There is therefore compelling morphologic and physiologic evidence for denervation changes in key upper airway dilator muscles in OSA.

The issue of denervation myopathy is of considerable potential importance with respect to in vivo upper airway contractile function. The in vitro human studies of Series et al^1–3 and Carrera et al,^5,6 which assessed contractility of musculus uvulae and genioglossus, did so using standard protocols of stimulation via wire electrodes of muscle strips in a bath. This approach will activate all muscle fibers, whether innervated or denervated. While electrically stimulated force generation was found to be similar in OSA and control muscle, the situation in vivo could be very different. In vivo, muscle contraction would be dependent on endogenous neurogenic activation—i.e., activation by the diseased nerves supplying the muscle rather than stimulating electrodes—which would be expected to be much less efficient in the presence of motor neuropathy, and in particular would preclude activation of denervated muscle fibers. Thus the in vitro contractility studies referred to above cannot be taken as conclusive evidence for normal muscle force generation by upper airway dilator muscles in vivo. The observations concerning denervation myopathy, together with the evidence for increased fatigability and dysfunctional mechanical coupling of upper airway dilators discussed above, provide compelling support for the concept of impaired in vivo upper airway dilator muscle contractile function in OSA.

There is histologic evidence for other upper airway myopathic changes in OSA. Petrof et al⁷ reported abnormal fiber morphology (central nucleation, fissured and moth-eaten appearance), inflammatory cell infiltrates, and increased connective tissue in sternohyoid and geniohyoid in the English bulldog. These changes were attributed to muscle overuse and injury due to upper airway loading,^7,10 as well as to eccentric contraction which may occur during airway occlusion.²⁶ Findings in human tissue studies have been discrepant, with some authors reporting grossly normal muscle histology^1,2,5 while others have described changes similar to those in the bulldog.^18,21,24,27 These disparate findings likely stem from differences in the muscles sampled, pathologic techniques, and patient characteristics, but also suggest that the extent of myopathic change varies between individuals. However, when present, pathologic changes of the nature described are undoubtedly associated with impaired force generation.

OSA is known to occur in clinical conditions involving pharyngeal neuromuscular dysfunction. Examples of this include the high prevalence of sleep apnea reported in Charcot-Marie-Tooth disease, a condition characterized by peripheral sensory and motor neuropathy.²⁸ OSA severity was correlated with the severity of changes in motor nerve conduction, suggesting that neurogenic upper airway myopathy may contribute to OSA. More recently, oculopharyngeal muscular dystrophy, which selectively affects the extraocular and pharyngeal muscles, was reported to be associated with OSA in the absence of obesity or abnormal upper airway morphology.²⁹ These observations suggest that a clinically identified myopathy can be associated with upper airway collapse which occurs uniquely during sleep, resulting in OSA.

The concept of an upper airway myopathy is entirely consistent with other current data on OSA pathophysiology. A fundamental observation is that, while upper airway dimensions are reduced during wakefulness in OSA, complete upper airway closure occurs only during sleep. Upper airway muscle EMG activity is increased during wakefulness in OSA, which is interpreted as indicating increased drive to, and force generation by, upper airway dilators to compensate for reduced anatomic dimensions.^30,31 When loss of waking tonic and phasic respiratory drive and attenuation of protective airway reflexes occurs at sleep onset, this is believed to result in inadequate dilator muscle compensation for reduced airway size, and complete collapse ensues.³¹

The concept of upper airway muscle dysfunction is entirely consistent with the other aspects of this schema. It is well established that EMG amplitude cannot necessarily be equated with muscle force generation, shortening, or tissue displacement.^32,33 Changes in excitation-contraction coupling can occur such that EMG amplitude can be maintained or even increase as force generation wanes in fatiguing or myopathic muscle.^32,33 Therefore, the augmented EMG activity described during wakefulness in OSA may reflect increased drive to upper airway muscles but does not necessarily indicate commensurately increased muscle force generation. Indeed, the fact that despite increased EMG activity, upper airway dimensions remain reduced during wakefulness in OSA strongly suggests inefficient muscle force generation (i.e., is consistent with myopathy).

With sleep onset, the reduction in drive to upper airway muscles could result in airway collapse, in part due to the fact that the inefficient force generation and tissue displacement by myopathic muscles are no longer adequate at the reduced level of drive to compensate for airway size and collapsibility. Through this mechanism, therefore, an upper airway myopathy could contribute to complete airway collapse, which occurs only during sleep. Consistent with this, at least one study³⁴ indicates that in mild forms of OSA, upper airway muscle training may reduce the number of respiratory events during sleep. Thus an increase in overall muscle force generating capacity and/or endurance due to training during wakefulness may lead to a beneficial impact on preservation of upper airway patency during sleep.

In summary, there is considerable evidence pointing to altered upper airway muscle contractile function in OSA and its role in the pathophysiology of this disorder. On the basis of the arguments presented here, we believe that to fail to pursue further investigation in this area, would be to ignore an important disease mechanism and perhaps even an eventual therapeutic target in this condition.