In Obstructive sleep apnea (OSA), the upper airway closes over and over again during sleep. This leads to disrupted sleep (waking up during the night), daytime sleepiness, and an increased risk for developing high blood pressure. Currently, the best treatment for obstructive sleep apnea is sleeping with a mask that continuously blows air into the nose (i.e. Continuous positive airway pressure [CPAP] treatment). While CPAP treatment stops the upper airway from closing in most people, many people have difficulty sleeping with the mask in place and therefore do not use the CPAP treatment. This research study is being conducted to learn whether using a combination of therapies (i.e. a sedative and oxygen therapy) will improve OSA severity by altering some of the traits that are responsible for the disorder.

Obstructive sleep apnea (OSA) is characterized by repetitive collapse or 'obstruction' of the pharyngeal airway during sleep. These obstructions result in repetitive hypopneas/apneas and intermittent hypoxia/hypercapnia, as well as surges in sympathetic activity. Such processes disturb normal sleep and impair neurocognitive function, often resulting in excessive daytime sleepiness and decreased quality of life. Furthermore, OSA is associated with cardiovascular morbidity and mortality, making OSA a major health concern.

Current evidence suggests that OSA pathogenesis involves the interactions of at least four physiological traits comprising 1) the pharyngeal anatomy and its propensity towards collapse 2) the ability of the upper airway dilator muscles to activate and reopen the airway during sleep (i.e. neuromuscular compensation), 3) the arousal threshold from sleep (i.e. the propensity for hypopneas/apneas to lead to arousal and fragmented sleep) and 4) the stability of the ventilatory feedback loop (i.e. loop gain). Continuous positive airway pressure (CPAP) is the most common treatment for OSA but it is often poorly tolerated; only ~50% of patients diagnosed with OSA continue therapy beyond 3 months. Given this limitation, alternative approaches have been tested and have generally focused on the use of oral appliances, surgery, and more recently pharmacological agents.

However, these alternate therapies, when used alone as monotherapy, rarely abolish OSA completely. This is not that surprising given that these treatments focus primarily on correcting only one trait and ignore the fact that the pathogenesis of OSA is multi-factorial. Thus the investigators hypothesize that some patients could be treated without CPAP if more than one trait is targeted (i.e., the investigators take a multi-factorial treatment approach). Such a multi-factorial approach is not unusual in Medicine. Many disorders such as diabetes, asthma, hypertension, cancer and congestive heart failure are treated with more than one medication or modality. In our view, giving CPAP to all OSA patients is like treating every diabetic with insulin, or every asthmatic with oral steroids - these treatments, like CPAP, are poorly tolerated and ignore the complexity of the underlying biology.

The investigators recently published a technique that measures the four traits using repeated 'drops' in CPAP levels during sleep. Each trait is measured in a way that allows model-based predictions of the presence/absence of OSA. With this technique the investigators demonstrated in a small group of CPAP-treated OSA subjects that decreasing the sensitivity of the ventilatory feedback loop (i.e. reducing loop gain) by approximately 50% with either acetazolamide or oxygen reduces the apnea/hypopnea index (AHI) by half. Interestingly, our model allowed us to make the prediction that if, in addition to an agent that reduces loop gain, the investigators also gave a drug that increases the arousal threshold by at least 25%, then the investigators could potentially abolish OSA (rather than just reduce its severity by 50%). This is of great interest given that the investigators already have shown than eszopiclone increases the arousal threshold by approximately 30% and is associated with an improvement in the AHI. However, to date there has been no study examining the combination of an agent that reduces loop gain (i.e. oxygen) with one that increases the arousal threshold (i.e. eszopiclone) as a treatment for OSA.

To determine the effect of combination therapy on each of the four traits and how they contribute to our model prediction of OSA, as well as on apnea severity. Specifically the investigators will assess:

1. The physiological traits responsible for OSA:

   1. Pharyngeal anatomy and its propensity towards collapse
   2. The ability of the upper airway dilator muscles to activate and reopen the airway during sleep (i.e. neuromuscular compensation)
   3. Arousal threshold from sleep (i.e. the propensity for hypopneas/apneas to lead to arousal and fragmented sleep).
   4. Stability of the ventilatory control system feedback loop (i.e. loop gain)

2. The severity of OSA (apnea-hypopnea index (AHI), percent of time with unstable breathing, sleep quality)

STUDY DESIGN:

A single-blinded randomized control design will be used. Initially, participants will be randomized to either the treatment or placebo arm where they will have both a clinical and research polysomnography (PSG); these initial PSGs constitute what will be referred to as VISIT 1 (see outcome measures). The purpose of the clinical PSG is to determine the severity of OSA (i.e. AHI). The research PSG will measure the 4 physiological OSA traits.

During the treatment arm, in both PSGs (i.e. clinical and research) participants will be given eszopiclone (3mg by mouth) to take before bed and be placed on oxygen throughout the night. During the placebo arm, subjects will be given a placebo to take before bed and placed on room air while they sleep. Participants will then have at least a 1-week washout period and cross over to the other arm of the study whereby the clinical and research PSG will be repeated; these studies constitute what will be referred to asVISIT 2 (see outcome measures).

Inclusion Criteria:

• Ages 18 - 79 years
• Documented OSA (AHI > 10 events/hr Non rapid eye movement sleep supine)
• If treated then, current CPAP use (>4 hrs CPAP/night for > 2 months)

Exclusion Criteria:

• Any uncontrolled medical condition
• Any other sleep disorder (Periodic leg movement syndrome, restless legs syndrome, insomnia, etc.)
• Use of medications known to affect sleep/arousal, breathing, or muscle physiology
• Allergy to lidocaine or Afrin
• Claustrophobia
• Alcohol consumption within 24 hours of PSG

During the placebo arm, one participant did not have OSA and another exhibited predominantly central sleep apnoea; both were excluded from the analysis.

The overall limitation is that we have only assessed the impact of combination therapy on one night. Thus, longitudinal studies will be needed to determine if the treatments remain effective.