1

Introduction

Sleep-disordered breathing results from a combination of factors affecting upper airway patency and the control of ventilation. Although the mechanisms underlying upper airway collapse are incompletely understood, a decline in pharyngeal neuromuscular activity during sleep appears to play a critical role. This knowledge has supported the notion that stimulation of upper airway muscles may represent a specific approach to the treatment of obstructive sleep apnea (OSA).

Although multiple upper airway dilator muscles play a role in maintaining upper airway patency during sleep, it has been widely accepted that the genioglossus (GG) muscle is one of the most important because activation advances the tongue. Investigators have considered tensor veli palatini function in animals and humans, but the majority of research related to electrical stimulation of upper airway musculature has described the tonic and reflexive activation of the GG muscle during wake and sleep. Consequently, methods have been explored to selectively stimulate upper airway dilator muscles, particularly the GG. This chapter will describe the foundation of research conducted in animals and humans that paved the way to provide upper airway stimulation as a treatment for OSA in adults.

2

Animal Studies

Miki et al. conducted the first animal experiments in this area and inserted needle electrodes periorally into the GG of spontaneously breathing dogs. A decrease in upper airway resistance was observed with progressively increasing stimulation frequencies up to 50 Hz. Schwartz et al. investigated the influence of bilateral supramaximal hypoglossal nerve (HGN) stimulation on upper airway mechanics. A graded increase in maximal inspiratory airflow (V _I max) was observed with increasing stimulation frequency. The improvement in V _I max could be attributed to a fall in upper airway collapsibility (reflected by a decrease in the critical closing pressure, or P _crit ), but the result was partially offset by a concomitant increase in upstream resistance. Interestingly, the observed fall in P _crit was of similar magnitude to that required for elimination of OSA in humans. In these experiments, the placement of the electrodes on the HGN was such that recruitment of different muscles (GG, styloglossus, hyoglossus, and intrinsic tongue muscles) could be obtained; stimulation of any or all of these might have contributed to the observed improvement in upper airway collapsibility.

Oliven et al. studied pressure–flow relationships of the upper airway during selective stimulation of the HGN in anesthetized dogs. Stimulation resulted in a significant decrease of upper airway resistance and an increase in P _crit and V _I max compared with controls. Eisele et al. conducted studies in the isolated feline upper airway to investigate how upper airway mechanics were altered by differential electrode placement along the HGN and the ansa cervicalis. From these data, it was concluded that a major decrease in upper airway collapsibility by HGN stimulation is dependent upon the activation of the GG and that electrode placement on the proximal segment of the HGN results in the largest improvement in V _I max. The importance of electrode placement and stimulation of specific upper airway muscles was also emphasized by Bishara et al. ; in spontaneously breathing dogs, selective stimulation of the GG with fine wire electrodes inserted into the muscle effectively reduced upper airway resistance and eliminated upper airway obstruction.

Goding et al. reported data on chronic stimulation of the HGN in dogs using bilateral cuff electrodes placed around the HGN. Stimulation started at 4 weeks after implantation and lasted for a period of 8 weeks at a ratio of 8 hours a day, 7 days a week. A major finding of this study, besides the improvement in peak upper airway flow during stimulation, was the absence of any damage to the nerve secondary to chronic stimulation. These results first suggested that long-term HGN stimulation in humans might be safe.

3

Initial Human Studies

The first attempts to improve upper airway patency in humans by electrical stimulation of upper airway muscles were made by Guilleminault et al. However, these first experiments were considered to be a failure. About 10 years later, Miki et al. reported their experience with submental electrical stimulation using an apnea-demand type stimulator in OSA patients. A decrease in Apnea Index and an improvement in sleep architecture could be documented. Later, these investigators used a portable airflow-demand type submental stimulator and reported similar findings. In this second study, however, it was acknowledged that only a partial improvement could be obtained with a decrease in the Apnea Index from 53.8 ± 7.0 to 27.3 ± 5.7 ( P < 0.05) and persistence of sleepiness.

These early favorable studies were promising; however, these results could not be reproduced by subsequent studies in other centers. A variety of methods of stimulation were attempted, but submental or intraoral stimulation, submental electrodes or fine wire electrodes placed into the neurovascular bundle, and transcutaneous electrical stimulation applied in the submental or infrahyoid regions all caused arousal from sleep before reaching the stimulation required to relieve upper airway obstruction.

Smith et al. also failed to obtain a significant improvement in airway patency without causing arousal when transcutaneous stimulation was applied through the submandibular region. However, when transoral fine wire electrodes were inserted into the GG, tongue protrusion and contralateral deviation (consistent with GG activation) were obtained during wakefulness. Posterior, rather than anterior, placement of the electrodes resulted in tongue retraction due to activation of the styloglossus and hyoglossus muscles.

Selective stimulation of upper airway muscles during sleep has been performed using transoral, intramuscular fine wire electrodes in patients with OSA. Stimulus bursts were first applied during single inspirations. A significant improvement in V _I max from 288.1 ± 176.2 mL/s to 501.4 ± 195 mL/s ( P < 0.001) was obtained during protrusor (GG) stimulation, whereas a significant decrease in V _I max was obtained with retractor (hyoglossus/styloglossus) stimulation. In these studies, stimulation did not cause arousal from sleep. Although no attempts were made to measure P _crit in these studies, the observed increase in V _I max during GG stimulation implied a fall in P _crit of approximately 4.8 cm H ₂ O. When stimulation was applied during consecutive inspirations, a significant decrease of sleep-disordered breathing episodes without worsening of sleep architecture could be demonstrated.